Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Using Low-Cost Radar Sensors and Action Cameras to Measure Inter-Vehicle Distances in Real-World Truck Platooning

Using Low-Cost Radar Sensors and Action Cameras to Measure Inter-Vehicle Distances in Real-World... Article Using Low-Cost Radar Sensors and Action Cameras to Measure Inter-Vehicle Distances in Real-World Truck Platooning 1 , 2 1 , 3 4 1 Markus Metallinos Log * , Thomas Thoresen , Maren H. R. Eitrheim , Tomas Levin and Trude Tørset Department of Civil and Environmental Engineering, Norwegian University of Science and Technology (NTNU), NO-7491 Trondheim, Norway Research Division Strategic Analyses and Joint Systems, Norwegian Defence Research Establishment (FFI), NO-2027 Kjeller, Norway Department of Humans and Automation, Institute for Energy Technology (IFE), NO-1751 Halden, Norway Norwegian Public Roads Administration (NPRA), NO-7031 Trondheim, Norway * Correspondence: markus.log@ntnu.no Abstract: Many modern vehicles collect inter-vehicle distance data from radar sensors as input to driver assistance systems. However, vehicle manufacturers often use proprietary algorithms to conceal the collected data, making them inaccessible to external individuals, such as researchers. Aftermarket sensors may circumvent this issue. This study investigated the use of low-cost radar sensors to determine inter-vehicle distances during real-world semi-automated truck platooning on two-way, two-lane rural roads. Radar data from the two follower trucks in a three-truck platoon were collected, synchronized and filtered. The sensors measured distance, relative velocity and signal-to-noise ratio. Dashboard camera footage was collected, coded and synchronized to the radar data, providing context about the driving situation, such as oncoming trucks, roundabouts and tunnels. The sensors had different configuration parameters, suggested by the supplier, to avoid signal interference. With parameters as chosen, sensor ranges, inferred from maximum distance measurements, were approximately 74 and 71 m. These values were almost on par with theoretical calculations. The sensors captured the preceding truck for 83–85% of the time where they had the preceding truck within range, and 95–96% of the time in tunnels. While roundabouts are problematic, the sensors are feasible for collecting inter-vehicle distance data during truck platooning. Citation: Log, M.M.; Thoresen, T.; Eitrheim, M.H.R.; Levin, T.; Tørset, T. Keywords: inter-vehicle distance measurements; radar sensor; action camera; field study; automated Using Low-Cost Radar Sensors and truck platooning; rural road; manual video coding; field-of-view; tunnels; roundabouts Action Cameras to Measure Inter-Vehicle Distances in Real-World Truck Platooning. Appl. Syst. Innov. 2023, 6, 55. https://doi.org/10.3390/ asi6030055 1. Introduction Truck platooning refers to the innovative concept of wirelessly linking trucks into Academic Editor: Igor Korobiichuk convoys using adaptive cruise control (ACC). Platoons consist of one lead truck and one or Received: 5 April 2023 more following trucks. Wireless communication may enable shorter inter-vehicle distances Revised: 27 April 2023 than those currently considered safe for manually driven trucks, which are constrained by Accepted: 4 May 2023 the reaction times of human drivers [1,2]. Shorter inter-vehicle distances lead to a reduction Published: 6 May 2023 in aerodynamic drag, which may enable fuel savings and reduced emissions. Moreover, tight, automated vehicle control may unlock improvements in safety and efficiency of the road traffic system [2]. Platooning may also benefit society at large, in terms of cheaper, safer and more streamlined road freight operations. However, platooning is yet to be Copyright: © 2023 by the authors. commercially deployed and there are many unanswered questions. Licensee MDPI, Basel, Switzerland. Field studies are often used to explore the technology and are organized either by This article is an open access article transport companies [3] or truck manufacturers directly [4], or through larger, publicly distributed under the terms and funded undertakings, such as the KONVOI [1] and ENSEMBLE projects [5]. The stud- conditions of the Creative Commons ies have typically been conducted on highways. Few studies, if any, have investigated Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ platooning on challenging two-way, two-lane roads with oncoming traffic and narrow 4.0/). tunnels. In Norway, for example, large parts of the road network are subject to such issues, Appl. Syst. Innov. 2023, 6, 55. https://doi.org/10.3390/asi6030055 https://www.mdpi.com/journal/asi Appl. Syst. Innov. 2023, 6, 55 2 of 29 on which more research is needed to establish the feasibility of truck platooning. Public roads authorities govern the design and operations of the road network and may thus be important facilitators of truck platooning. As Norway is a small automotive market with conditions for automated vehicles, the Norwegian Public Roads Administration (NPRA) have taken a proactive role in trialing advanced transportation technologies, exemplified by the Borealis testbed [6] and the ongoing MODI project [7], which aims to demonstrate automated trucking between the Netherlands and Norway within 2026. In general, the shorter inter-vehicle distances are between platooning trucks, the greater are the resulting benefits in terms of fuel savings and potential road capacity improvements. Thus, the greater the stability of inter-vehicle distances over time, the more beneficial platooning will be [8]. However, during real-world driving, combinations of external traffic, road alignment and truck weight differences will influence the inter-vehicle distances, and thus also the extent of benefits unlocked [9,10]. Moreover, the extent to which truck platoons impact surrounding traffic depends partly on their total length, which is influenced by inter-vehicle distances. By implication, inter-vehicle distance data will be important for public roads authorities when regulating truck platooning, such as when deciding on which road sections platoons should be allowed, and the maximum number of trucks which can platoon together. However, such data may not be easily accessible, and even if they were, it is unclear how they should be contextualized and analyzed. In the field studies organized through the large, aforementioned platooning projects, truck manufacturers and transport companies have typically facilitated and allowed for the collection of inter-vehicle distance measurements using integrated vehicle sensors. While integrated distance sensors could also be cameras and lidars [11], radar sensors are most often used, as they are affordable, computationally simple and robust under adverse light and weather conditions [11–15]. This is despite issues with clutter [16] and ghosting [17], referring to unwanted signals that distort and interfere with the desired detection, and the detection of non-existent targets which are difficult to distinguish from real ones. If allowed, using data from integrated vehicle sensors is very convenient, as they are already collected as real-time inputs to platooning control systems, removing the need for using aftermarket distance sensors. However, manufacturers often use proprietary algorithms for data processing, defining message codes [18] to encrypt collected data. This makes external individuals unable to access them, unless authorized to do so. Smaller truck platooning field studies may not have the benefit of truck manufacturers participating as partners. Some may also seek to verify data from manufacturers, for which independent methods for collecting such data would be useful. High-precision global navigation satellite system (GNSS) receivers located in succes- sive vehicles could theoretically be used for this purpose [19,20]. Some areas, however, have road tunnels and topographical features where GNSS-based data collection methods would be subject to signal blockages [12,20]. External aftermarket radar sensors may circumvent the problem, provided they are adequately practical and accurate. Truck platooning re- search is often publicly funded, so solutions should preferably be low-cost. Since the output of aftermarket radar sensors would not be used for operative vehicle control, they do not need to be as capable, nor provide the same level of reliability as automotive-grade distance sensors, both of which drive cost and complexity. They may also be more flexible, in terms of allowing for custom placement and user adaptations. Many researchers have focused on perception, functional safety and operative control for truck platoons [21]. These studies often include cameras and radars. However, for studying the effects of truck platooning from the standpoint of roads authorities, these methods are more computationally complex than they need to be, and simple methods for estimating inter-vehicle distances from truck platooning would be useful. The current study investigated the feasibility of using Anteral universal radar (uRAD) sensors for Raspberry Pi [22] to measure inter-vehicle distances in a truck platoon on rural roads. This application represents a novel use case for this type of sensor. We propose a multi-faceted approach for collecting inter-vehicle distance data from truck Appl. Syst. Innov. 2023, 6, x FOR PEER REVIEW 3 of 29 on rural roads. This application represents a novel use case for this type of sensor. We propose a multi-faceted approach for collecting inter-vehicle distance data from truck pla- tooning field trials. It aims to provide technical details and best-practices on data collec- tion, synchronization, filtering and analysis. Dashboard cameras in each truck filmed the driving scene. This footage was used to log the timestamps of specific, recurring events, allowing for exploring sensor operation in different driving conditions. Appl. Syst. Innov. 2023, 6, 55 3 of 29 The paper addresses the research question: How can low-cost radar sensors and ac- tion cameras be used to investigate inter-vehicle distances in real-world truck platooning? platooning field trials. It aims to provide technical details and best-practices on data 2. Materials and Methods collection, synchronization, filtering and analysis. Dashboard cameras in each truck filmed This section provides details on the data collection set-up, the equipment used, and the driving scene. This footage was used to log the timestamps of specific, recurring events, the procedures for synchronization, video coding and radar data processing. allowing for exploring sensor operation in different driving conditions. The paper addresses the research question: How can low-cost radar sensors and action 2.1. Data Dollection Set-Up cameras be used to investigate inter-vehicle distances in real-world truck platooning? A truck platooning field trial was undertaken on public rural roads in northern Nor- way in the fall of 2020. This is the first study of its kind, and was also reported in [23]. 2. Materials and Methods Three drivers operated three semi-trailer trucks along a 380 km two-way, two-lane road This section provides details on the data collection set-up, the equipment used, and stretch traversing a mountainous, coastal area. The trucks were numbered 1, 2 and 3, the procedures for synchronization, video coding and radar data processing. based on the main truck order configuration. One section was traversed repeatedly with different orders. A prototype ACC system was installed, enabling the trucks to operate as 2.1. Data Dollection Set-Up a platoon when detecting a preceding truck. Data from integrated cameras and radars A truck platooning field trial was undertaken on public rural roads in northern Norway were unavailable. Aftermarket equipment was used to collect data over 7 h of driving. in the fall of 2020. This is the first study of its kind, and was also reported in [23]. Three While longitudinal control was automated, the drivers operated the wheel manually, drivers operated three semi-trailer trucks along a 380 km two-way, two-lane road stretch placing the field study at the Society of Automotive Engineers (SAE) Level 1 [24]. All traversing a mountainous, coastal area. The trucks were numbered 1, 2 and 3, based on the trucks had 500 horsepower. Trucks 1 and 2 had equal weights, while truck 3 was lighter. main truck order configuration. One section was traversed repeatedly with different orders. (41 and 27.5 metric tons), reflecting the scenario of trucks encountering each other to pla- A prototype ACC system was installed, enabling the trucks to operate as a platoon when toon during real-world operations, unharmonized with respect to weight. Consequently, detecting a preceding truck. Data from integrated cameras and radars were unavailable. inter-vehicle distances fluctuated, as the platoon often struggled to remain collected on Aftermarket equipment was used to collect data over 7 h of driving. the winding road. Twenty-three tunnels and eleven roundabouts were traversed. The While longitudinal control was automated, the drivers operated the wheel manually, most prevalent speed limits were 80 and 60 km/h, during which inter-vehicle distances at placing the field study at the Society of Automotive Engineers (SAE) Level 1 [24]. All trucks 3-second time gaps were 50–70 m, i.e., comparable to manual driving. Figure 1 illustrates had 500 horsepower. Trucks 1 and 2 had equal weights, while truck 3 was lighter. (41 and the set-up. 27.5 metric tons), reflecting the scenario of trucks encountering each other to platoon during Herein, preceding truck is a term used for the truck located in front of the truck in real-world operations, unharmonized with respect to weight. Consequently, inter-vehicle question. Depending on the context, it may refer to either the lead truck or the middle distances fluctuated, as the platoon often struggled to remain collected on the winding truck. In truck order 1-2-3, truck 1 precedes truck 2, and truck 2 precedes truck 3. The road. Twenty-three tunnels and eleven roundabouts were traversed. The most prevalent terms leader or leading truck are only used for the truck located at the front of the platoon, speed limits were 80 and 60 km/h, during which inter-vehicle distances at 3-s time gaps while followers refer to both the middle and last truck together. were 50–70 m, i.e., comparable to manual driving. Figure 1 illustrates the set-up. Figure 1. Truck platooning set-up. Radar sensors were mounted on windshields of following trucks. Figure 1. Truck platooning set-up. Radar sensors were mounted on windshields of following trucks. The two rearmost trucks were identically instrumented. The Raspberry Pi, with the Herein, preceding truck is a term used for the truck located in front of the truck in uRAD sensor att ached, was fixed to the inside of the windshield using a suction mount question. Depending on the context, it may refer to either the lead truck or the middle with a flexible arm. The mount did not interfere with the field-of-view of the sensors. A truck. In truck order 1-2-3, truck 1 precedes truck 2, and truck 2 precedes truck 3. The terms leader or leading truck are only used for the truck located at the front of the platoon, while followers refer to both the middle and last truck together. The two rearmost trucks were identically instrumented. The Raspberry Pi, with the uRAD sensor attached, was fixed to the inside of the windshield using a suction mount with a flexible arm. The mount did not interfere with the field-of-view of the sensors. A portable monitor was used to administer radar logging. Sensors were placed at slightly different heights in each truck due to interior constraints, cf. Appendix A.1. Two GoPro video cameras were also mounted: A windshield-mounted dashboard camera filmed the driving scene, and another camera filmed the interior. Footage from Appl. Syst. Innov. 2023, 6, x FOR PEER REVIEW 4 of 29 portable monitor was used to administer radar logging. Sensors were placed at slightly different heights in each truck due to interior constraints, cf. Appendix A.1. Appl. Syst. Innov. 2023, 6, 55 4 of 29 Two GoPro video cameras were also mounted: A windshield-mounted dashboard camera filmed the driving scene, and another camera filmed the interior. Footage from the latt er was only used for synchronizing radar data to the dashboard footage. The study the was a latter pprov was ed by the Norw only used for synchr egian Centre for onizing radar Research Da data to the ta (4 dashboar 57013). P d footage. articipan The ts ag study reed to being recorded, and all videos and audio were handled and stored confidentially. was approved by the Norwegian Centre for Research Data (457013). Participants agreed to The equipment in each truck was started and stopped in succession when trucks were being recorded, and all videos and audio were handled and stored confidentially. parked, but The equipment was left on during short brea in each truck was started ks. None and stopped of the Raspb in succession erry Pi when micro trp ucks rocessors were parked, had internet connecti but was left on on, so during they di short d not a breaks. dhere to l None ofothe cal ti Raspberry me. An equi Pi micr pment sta oprocessors rt-up pro- had internet connection, so they did not adhere to local time. An equipment start-up procedure cedure was devised which allowed for post-hoc synchronization of videos and radar data. was devised which allowed for post-hoc synchronization of videos and radar data. For For each truck, all cameras were started before starting radar logging. When starting each each truck, all cameras were started before starting radar logging. When starting each camera, the Emerald Sequoia Time smartphone application was presented. Using Net- camera, the Emerald Sequoia Time smartphone application was presented. Using Network work Time Protocol (NTP) servers, which synchronizes computer clocks over the internet Time Protocol (NTP) servers, which synchronizes computer clocks over the internet [25], [25], this application provides more accurate times than those typically provided by inter- this application provides more accurate times than those typically provided by internal nal clocks [26]. The application shows local time, and, when cellular reception is available, clocks [26]. The application shows local time, and, when cellular reception is available, it it calculates deviations from NTP time. The mean offset was 0.08 s, i.e., negligible. calculates deviations from NTP time. The mean offset was 0.08 s, i.e., negligible. GNSS data were collected from VBOX Sport loggers and a fleet management system GNSS data were collected from VBOX Sport loggers and a fleet management system (FMS), in an effort to compute inter-vehicle distances to validate the radar data. A script (FMS), in an effort to compute inter-vehicle distances to validate the radar data. A script was writt en to interpolate timestamps and calculate distances between GNSS locations was written to interpolate timestamps and calculate distances between GNSS locations from each truck. The loggers were supposed to activate automatically [27], but this func- from each truck. The loggers were supposed to activate automatically [27], but this func- tionality occasionally failed. GNSS files were also extracted from the FMS, and all files tionality occasionally failed. GNSS files were also extracted from the FMS, and all files were visualized in QGIS. Both systems experienced outages in tunnels. While the FMS were visualized in QGIS. Both systems experienced outages in tunnels. While the FMS had good positioning accuracy outside tunnels, its update rate was too low, and loggings had good positioning accuracy outside tunnels, its update rate was too low, and loggings were not always synchronized across the trucks. VBOX data which did get collected had were not always synchronized across the trucks. VBOX data which did get collected had frequent outages, and timestamps were often erroneous, placing trucks in incorrect order. frequent outages, and timestamps were often erroneous, placing trucks in incorrect order. This highlights the utility of radar in estimating inter-vehicle distances in such areas. This highlights the utility of radar in estimating inter-vehicle distances in such areas. 2.2. Radar Sensors 2.2. Radar Sensors Frequency modulated continuous waves (FMCWs) are radar waveforms often used Frequency modulated continuous waves (FMCWs) are radar waveforms often used to measure distances in automotive applications [13,28]. Anteral uRAD radar sensors for to measure distances in automotive applications [13,28]. Anteral uRAD radar sensors for Raspberry Pi were tested here, shown in Figure 2. These are 24 GHz FMCW radar sensors Raspberry Pi were tested here, shown in Figure 2. These are 24 GHz FMCW radar sensors which connect conveniently as extension boards to Raspberry Pi microprocessors [29]. which connect conveniently as extension boards to Raspberry Pi microprocessors [29]. Such Such microprocessors run a user-friendly operating system and can interface with pur- microprocessors run a user-friendly operating system and can interface with purpose- pose-built components. The automotive industry is increasingly using 77 GHz radar sen- built components. The automotive industry is increasingly using 77 GHz radar sensors, sors, allowing for increased range resolution and accuracy [30]. These sensors can bett er allowing for increased range resolution and accuracy [30]. These sensors can better separate separate closely spaced objects, and can be packaged in a smaller form factor. However, closely spaced objects, and can be packaged in a smaller form factor. However, 24 GHz 24 GHz sensors are less expensive, and automotive-grade 77 GHz sensors which could be sensors are less expensive, and automotive-grade 77 GHz sensors which could be operated operated from Raspberry Pi microprocessors were not available when procuring the from Raspberry Pi microprocessors were not available when procuring the equipment. equipment. Thus, testing the proposed methodology using cheaper 24 GHz sensors was Thus, testing the proposed methodology using cheaper 24 GHz sensors was considered considered reasonable. While many 77 GHz FMCW radar sensors are more range-capable reasonable. While many 77 GHz FMCW radar sensors are more range-capable and have and have wider fields-of view, some 77 GHz radars, e.g., in [31], have shorter ranges than wider fields-of view, some 77 GHz radars, e.g., in [31], have shorter ranges than the the uRAD sensors. uRAD sensors. Figure 2. Radar sensor outline [22]. Appl. Syst. Innov. 2023, 6, 55 5 of 29 The uRAD sensors have a theoretical distance range approaching 100 m, 30 fields-of- view both horizontally and vertically, and are able to detect up to five objects simultaneously. For each object, distance, radial velocity and reflected power (SNR, i.e., signal-to-noise ratio) are registered. In the context of truck platooning, both the sensor and the desired objects are in motion; so relative velocities are detected. The velocity range (0.2 to 75 m per second) is within the range of values encountered in road traffic. The use case and intended sensor placement were described to the supplier, which expected the application to be feasible, but did note that similar tests had never been carried out before. The sensor has been used for other purposes, the most relevant of which is as a stationary speed sensor [29,32,33]. However, transferability from the cited studies is limited, due to different use cases and configurations. Eight parameters were used to configure the sensors. These are detailed in Appendix A.1, alongside pre-trial testing of the mounting set-up and parameters. The most important con- siderations are detailed here. Firstly, the radar mode details the waveforms transmitted by the sensor. Triangular waves were chosen, which maximized the sensor range and the update rate for outputs (9–13 Hz) from the radar script. This mode also allowed for subsequent data filtering based on relative velocity. Maximizing range was important for capturing data even when trucks were located far apart, as the trucks were expected to drive with human-level gap sizes (2–3 s) at distances approaching the upper distance range. Adverse road geometry would presumably also lead to safer driving at larger gap sizes. Maximizing the update rate was seen as beneficial for obtaining as many measurements as possible. The update rate of the uRAD radar sensor is comparable to the 77 GHz sensors showcased in [14]. Secondly, number of targets detected (N ) and the detection distance (R ) were tar max maximized, to capture the most data, and to enable filtering of unwanted detections later. Thirdly, moving Target Indicator (MTI) was activated, for including data only from objects with motion relative to the sensor. The supplier stated that it would only eliminate objects which were absolutely static, such as detections of the windshield. The preceding truck would still be registered, even when moving at the same velocity as the sensor. Fourthly, for each truck, different values for ramp start frequency (f ) and the duration of each wave ramp (N ) were used for each sensor to avoid interference. Since each sensor had different N and f values, their theoretical maximum distance ranges also s 0 differed, at 75.0 and 73.1 m, for trucks 2 and 3, respectively, based on Equation A3 in Appendix A.1. These values are in line with 70–75 m estimates from the supplier. For comparison, automotive radar sensors typically have ranges of 30–150 m [30,34]. The sensors had a stated distance accuracy of 0.3%, corresponding to a 0.23 m deviation at 75 m, which is considered sufficient for the current use case. Table A3 in Appendix A.1 shows all parameters which were used. 2.3. Video Footage, Synchronization and Manual Video Coding Dashboard footage was recorded for exploring the radar data as a function of the driving scene. Without the videos, this would not have been possible. First, footage had to be coded, i.e., timestamps had to be established for relevant events in the footage. This is different from the more computationally complex process of semantic segmentation used in computer vision, which involves categorizing relevant objects in the scene, often using bounding boxes [31]. Video footage was synchronized and aligned to local time using BORIS, i.e., Behavioral Observation Research Interactive Software [35]. BORIS is a free, open-source video coding program. Each BORIS observation contained videos from the same truck and driving stretch. By checking time differences as displayed by the phone application at the start of each recording, time offsets were established, achieving near-perfect synchronization. The date and time of each observation was defined as the local time shown by the phone application to the longest video file, as the recording was started. This ensured that all videos from the trucks were aligned to local time during the field trial. Appl. Syst. Innov. 2023, 6, x FOR PEER REVIEW 6 of 29 Appl. Syst. Innov. 2023, 6, 55 6 of 29 application to the longest video file, as the recording was started. This ensured that all videos from the trucks were aligned to local time during the field trial. Events in BORIS were defined using the ethnogram, and were either point or state Events in BORIS were defined using the ethnogram, and were either point or state events: Point events had no duration (i.e., having only one timestamp), while state events events: Point events had no duration (i.e., having only one timestamp), while state events did (i.e., having both start and end timestamps). Video coding was carried out while play- did (i.e., having both start and end timestamps). Video coding was carried out while ing videos at 2–4 times normal speed, depending on driving scene complexity. Events playing videos at 2–4 times normal speed, depending on driving scene complexity. Events were coded by the first author, ensuring consistency. Videos codes were subsequently re- were coded by the first author, ensuring consistency. Videos codes were subsequently viewed by the third author. Onwards, italics are used to refer to the video codes, as illus- reviewed by the third author. Onwards, italics are used to refer to the video codes, as trated in Figure 3. Radar initiations (Radar logging) and oncoming traffic were defined as illustrated in Figure 3. Radar initiations (Radar logging) and oncoming traffic were defined point events. Amongst oncoming vehicles, only Trucks seemed to affect inter-vehicle dis- as point events. Amongst oncoming vehicles, only Trucks seemed to affect inter-vehicle tances during platooning. Video footage showed that when encountering large trucks on distances during platooning. Video footage showed that when encountering large trucks on narrow road segments, the lead truck often reduced its speed, causing speed reductions narrow road segments, the lead truck often reduced its speed, causing speed reductions also also for the followers and a contraction in inter-vehicle distances. Truck order codes indi- for the followers and a contraction in inter-vehicle distances. Truck order codes indicate cate which periods the instrumented trucks collected relevant inter-vehicle distance data which periods the instrumented trucks collected relevant inter-vehicle distance data within within the platoon, as opposed to the periods where they served as platoon leaders, col- the platoon, as opposed to the periods where they served as platoon leaders, collecting lecting irrelevant data preceding the platoon. irrelevant data preceding the platoon. Figure 3. Overview of BORIS video codes. Figure 3. Overview of BORIS video codes. The Driving video code in Figure 3 includes all conditions encountered, i.e., including The Driving video code in Figure 3 includes all conditions encountered, i.e., includ- tunnels and roundabouts, thus showcasing diverse, complex driving segments. See Table ing tunnels and roundabouts, thus showcasing diverse, complex driving segments. See A5 in Appendix A.3 for its definition. Tunnel and Roundabout video codes denote scenarios Table A5 in Appendix A.2 for its definition. Tunnel and Roundabout video codes denote of particular interest. In tunnels, inter-vehicle distances between the trucks cannot be de- scenarios of particular interest. In tunnels, inter-vehicle distances between the trucks can- termined using GNSS-based methods, and it is also unclear how tunnels affect operating not be determined using GNSS-based methods, and it is also unclear how tunnels affect conditions for the radar sensors. Roundabouts are demarcated areas (i.e., they are simple operating conditions for the radar sensors. Roundabouts are demarcated areas (i.e., they to code from video footage) with small horizontal radii, which can illustrate effects of road are simple to code from video footage) with small horizontal radii, which can illustrate ef curv fects ature of r on oad rad curvatur ar operat e on ion when radar operation the preceding when truck turn the preceding s. All e tr vuck ents w turns. ere co All ded events sep- wer arately e coded for eseparately ach truck. for Tun each nels w truck. ere co Tunnels ded from the were coded moment when the front of from the moment whenthe truc the front k in question entered the tunnel, to when the front of the truck left the tunnel. The same of the truck in question entered the tunnel, to when the front of the truck left the tunnel. The principle w same principle as used for rounda was used for bouts, i.e., codi roundabouts, ng the i.e., coding moment when the front of each truck the moment when the front of each entered truck and e enter xited the circ ed and exited ulating the cir are culating a. area. Events were defined for visual inspection of the distance to (visibility of) the preceding Events were defined for visual inspection of the distance to (visibility of ) the preced- truck, as the trucks were at times located far apart. The goal was to remove data from ing truck, as the trucks were at times located far apart. The goal was to remove data from periods when the preceding truck was difficult or impossible for the sensor to detect, due periods when the preceding truck was difficult or impossible for the sensor to detect, due to the driving situation. This occurred in two scenarios. Firstly, it occurred in sharp turns, to the driving situation. This occurred in two scenarios. Firstly, it occurred in sharp turns, where the preceding truck would disappear from radar field-of-view. Dashboard cameras where the preceding truck would disappear from radar field-of-view. Dashboard cameras had larger horizontal fields-of-view than radar sensors, so when no preceding vehicle had larger horizontal fields-of-view than radar sensors, so when no preceding vehicle ap- appeared on camera, the radar would also not detect it. Secondly, it occurred when the peared on camera, the radar would also not detect it. Secondly, it occurred when the trucks drove far apart. The distance range was shorter for radar than for the dashboard trucks drove far apart. The distance range was shorter for radar than for the dashboard camera, which was only constrained by line-of-sight. Both scenarios were coded as Likely Appl. Syst. Innov. 2023, 6, 55 7 of 29 beyond radar range (LBRR). Conversely, Likely within radar range (LWRR) denotes driving periods when relevant radar data likely could have been collected. The manual nature of this process introduces some limitations. Periods when the preceding truck was actually LWRR may have been coded as LBRR, and vice versa. Tran- sitions between these codes may also occur at different distances. Albeit imperfect, this categorization is preferable versus including all radar data, even when the preceding truck was located far beyond radar range. For the far-apart scenario, centerline road markings initially aided the visual estimation. On rural roads with 80–90 km/h speed limits, these markings have standardized lengths and gaps totaling 12 m, which repeat continuously, cf. pp. 22 in the Norwegian road marking design manual [36]. After having coded LWRR and LBRR using road markings for some time, the remaining dashboard camera footage was coded without conscious reference to the road markings. It was also attempted to use pixel counts of the preceding truck for this purpose, but doing so at large scale was unsuccessful. An overview of the data collection and processing steps is provided in Table 1. See Table A5 in Appendix A.2 for examples of video codes. Table 1. Overview of the data collection and processing steps. Step Context Description Start GoPro-cameras successively, while, for each camera, presenting local time on phone screen. Equipment start-up and logging 2 Start radar logging script while producing loud verbal cue. 3 Data collection Platoon driving. Stop GoPro camera recordings successively. Stop radar 4 Equipment logging stop logging. 5 Data transfer Import GoPro video files and raw radar files to computer. For each truck: Synchronize GoPro video footage in BORIS, using offset values. Synchronization is based on the Synchronize GoPro videos with each 6 difference between local time presented to each camera other upon starting the recordings, and fine-tuned using recorded audio. Define Date and time in BORIS observation equal to the local time shown to the reference camera (i.e., the longest video 7 Synchronize GoPro videos to local time file) by the phone application when the reference recording was started. Code Radar logging based on visual and verbal cues from 8 Video coding interior camera. Code remaining events from dashboard camera footage. 9 Export events list for each observation to spreadsheets. Apply datetime shift to radar timestamps based on Date and Synchronize radar data to local time 10 time for each BORIS observation to match them with Radar logging events. 11 Radar data curation Radar data were curated using six filters. 2.4. Radar Data Processing Video coding events were exported from BORIS as spreadsheets, and the Date and time from the corresponding observation was added to the timestamp of each instance of the Radar logging video code. This assigned local time to the instance when radar logging was started, and served as basis for synchronizing video codes and radar data in Python. Filters were needed to extract only the inter-vehicle distances between the platooning trucks. Filtering aimed at removing data from periods when the trucks were not driving (i), data which did not correspond to the preceding truck (ii), noise (iii), and finally, data from periods when the preceding truck was outside sensor range (iv). Appl. Syst. Innov. 2023, 6, 55 8 of 29 Appl. Syst. Innov. 2023, 6, x FOR PEER REVIEW 8 of 29 Timestamps, distance (m), relative velocity (km/h) and signal-to-noise ratio (decibel, (i), data which did not correspond to the preceding truck (ii), noise (iii), and finally, data dB) were logged, for up to five simultaneously detected targets. Positive relative velocities from periods when the preceding truck was outside sensor range (iv). corresponded to targets receding from the radar, and negative relative velocities corre- Timestamps, distance (m), relative velocity (km/h) and signal-to-noise ratio (decibel, sponded to approaching targets. For the curated radar dataset, an example of the former dB) were logged, for up to five simultaneously detected targets. Positive relative velocities would be the preceding truck accelerating away from the truck in question. Conversely, corresponded to targets receding from the radar, and negative relative velocities corre- the preceding truck decelerating would be an example of the latter. SNR denotes the ratio sponded to approaching targets. For the curated radar dataset, an example of the former of the signal power to the noise power. Larger and more reflective objects will produce would be the preceding truck accelerating away from the truck in question. Conversely, measurements with higher SNR values. The radar data were curated using successive the preceding truck decelerating would be an example of the latt er. SNR denotes the ratio filters, cf. Table 2. The following paragraph outlines the details and purpose of each filter. of the signal power to the noise power. Larger and more reflective objects will produce measurements with higher SNR values. The radar data were curated using successive fil- Table 2. Overview of radar data filters. ters, cf. Table 2. The following paragraph outlines the details and purpose of each filter. Filter Description Table 2. Overview of radar data filters. 1 Driving and following Filter Description 2 Relative velocity within 30 km/h 1 Driving and following 3 Signal-to-noise ratio < 15 dB 2 Relative velocity within ±30 km/h 4 Target selection 3 Signal-to-noise ratio < 15 dB 4 T 5 arget selection Downsampling 1 Hz 5 Downsampling 1 Hz 6 Likely within radar range (LWRR) 6 Likely within radar range (LWRR) First, the Driving video code was used to remove data collected during irrelevant First, the Driving video code was used to remove data collected during irrelevant pe- periods. It discards data from Break periods, so only data from Driving periods remain. riods. It discards data from Break periods, so only data from Driving periods remain. Sim- Simultaneously, truck order codes were used to exclude radar data collected during periods ultaneously, truck order codes were used to exclude radar data collected during periods when each truck served as platoon leader. Specifically, radar data from truck 2 stem from when each truck served as platoon leader. Specifically, radar data from truck 2 stem from the driving periods with truck orders 1–2–3, 1–3–2 and 3–2–1, while radar data from truck the driving periods with truck orders 1–2–3, 1–3–2 and 3–2–1, while radar data from truck 3 stem from periods coded as 1–2–3, 1–3–2 and 2–1–3. Data were discarded from periods 3 stem from periods coded as 1–2–3, 1–3–2 and 2–1–3. Data were discarded from periods where external vehicles (Other vehicle) preceded each respective truck. where external vehicles (Other vehicle) preceded each respective truck. Filters for relative velocity and SNR (filters 2 and 3, respectively), were used to clean Filters for relative velocity and SNR (filters 2 and 3, respectively), were used to clean the remaining data. Relative velocities were explored in histograms, shown in Figure 4a. the remaining data. Relative velocities were explored in histograms, shown in Figure 4a. Some relative velocity bins were far more frequent than others, giving the dataset large Some relative velocity bins were far more frequent than others, giving the dataset large dynamic range. The vertical axis is logarithmic, magnifying bins with few measurements. dynamic range. The vertical axis is logarithmic, magnifying bins with few measurements. (a) (b) Figure 4. (a) Relative velocity histograms. (b) Distance histograms with relative velocity filters. Appl. Syst. Innov. 2023, 6, 55 9 of 29 The histograms show two data spikes. Relative to a forward-facing sensor mounted in a moving truck, stationary surroundings have negative speed comparable to the speed of the truck. Assuming traffic moves at the speed limit, oncoming vehicles are measured with negative relative velocities at twice the speed limit. Similarly, measurements from the preceding truck have relative velocities fluctuating around 0 km/h. These two clusters of relative velocities appear as vertical spikes in Figure 4a. The cluster from the preceding truck was the largest. This is as expected in car-following situations, which necessitate continuous acceleration and deceleration [37]. The smaller clusters in Figure 4a had an order of magnitude fewer measurements, all of which had relative velocities in the 30 to 160 km/h range. They included static objects, oncoming traffic and measurement noise. Thus, for the relative velocity filter (filter 2), choosing 30 km/h as the lower threshold was natural, placing it at the local minimum between the spikes. Similarly, setting the upper threshold at +30 km/h made it so the entire top spike was included, while minimizing the inclusion of measurements from the noise floor. As shown in Figure 4b, the radar data were also subjected to different relative veloc- ity filters, starting at 40 km/h, successively constricting by 10 km/h steps until the narrowest filter of 10 km/h. The color of each filter reflects remaining data points after that filter has been applied. For instance, remaining data after relative velocity filtering at 30 km/h are shown in red. The 40 km/h filter left a spike from 10–20 m for truck 3. The 30 km/h filter removed most of the spike, and subsequent filter constriction did not cause notable differences. Thus, 30 km/h was chosen, striking a balance between retaining most measurements corresponding to the preceding trucks and minimizing unwanted ones (oncoming and stationary objects), while including situations with sudden braking and acceleration in the platoon, which are perhaps the most interesting ones from a safety and fuel savings perspective. The radar sensors were listed as having a velocity accuracy of 0.25 m per second. At 30 km/h cut-offs, this amounts to a possible deviation of 3%, which is considered acceptable. Still, as shown in Figure 4b, relative velocity filtering did not remove the leftmost spikes of measurements at distances too short to represent the preceding truck (0–5 m). Since the platoon drove with 2–3 s gaps, these do not represent the preceding truck. Calculations using Equation (A1), based on the 30 vertical field-of-view of the sensor, show that the road can be detected at 8–8.5 m forwards, meaning further away than the spikes. Thus, they are most likely clutter due to roadside detections, such as such as rock faces, tunnel walls, guardrails and signposts. Compared to the large, reflective rear walls of preceding trucks, such measurements should presumably be noisy, i.e., have small SNR values. Conversely, if these measurements did originate from the preceding truck, they should have had large accompanying SNR values. SNR values at short distances were indeed found to be small, and filtering for SNR < 15 dB was successful in removing them. As shown in Figure 5a, 15 dB filtering fell at a local minimum or saddle point between two distinct SNR data spikes. The data were also subjected to different SNR filters, illustrated in Figure 5b, but the filter was not constricted further, as doing so caused removal of data points with distance values around 40 m, likely corresponding to the preceding truck. Since the radar sensors were able to measure up to 5 detections simultaneously, the next filter (filter 4) involved selecting only one desired target in multi-target instances: The one most likely corresponding to the preceding truck. The distance value of each detection was compared with the average distance values of the previous 10 measurements (moving average). The detection with the smallest difference was chosen. However, occasional single-object detections had distances which were quite different from the general trend. In such cases, the algorithm had no choice but to select the only detection available. This produced spikes or drops in distance which affected the moving average. This problem was subsequently minimized by filter 5, which downsampled the data to 1 Hz (one measurement per second) by averaging all distance measurements within each second. This temporal resolution was considered sufficient, and also allowed for direct coupling between curated radar data and event durations from video footage. Appl. Syst. Innov. 2023, 6, 55 10 of 29 Appl. Syst. Innov. 2023, 6, x FOR PEER REVIEW 10 of 29 (a) (b) Figure 5. (a) SNR histograms after relative velocity filtering. (b) Distance histograms with different Figure 5. (a) SNR histograms after relative velocity filtering. (b) Distance histograms with different SNR threshold (after relative velocity filtering). SNR threshold (after relative velocity filtering). Since the radar sensors were able to measure up to 5 detections simultaneously, the Inspecting plots of distance versus time for downsampled radar data revealed the next filter (filter 4) involved selecting only one desired target in multi-target instances: The presence of sporadic periods entirely void of points, and also to periods when points were one most likely corresponding to the preceding truck. The distance value of each detection scattered (i.e., having varying distance values following no obvious trend). To understand was compared with the average distance values of the previous 10 measurements (moving these detections, video footage was coded for LWRR and LBRR. Filter 6 used these codes to average). The detection with the smallest difference was chosen. However, occasional sin- include only data collected in LWRR periods, and to exclude data collected during LBRR gle-object detections had distances which were quite different from the general trend. In periods. Cameras malfunctioned at times, during which LWRR and LBRR could not be such cases, the algorithm had no choice but to select the only detection available. This coded. Associated radar data were discarded, ensuring methodological consistency. produced spikes or drops in distance which affected the moving average. This problem In brief, filters 1 and 6 were based on manual video codes, while filters 2 and 3 were was subsequently minimized by filter 5, which downsampled the data to 1 Hz (one meas- based on recorded radar metrics, making them the most interesting ones in terms of radar urement per second) by averaging all distance measurements within each second. This operation. Filters 4 and 5 were computational heuristics. See Appendix A.3 for more details. temporal resolution was considered sufficient, and also allowed for direct coupling be- 3. Results and Discussion tween curated radar data and event durations from video footage. Inspecting plots of distanc This section explores effects e ve of rsus the time filtering for downsampled radar process, before discussing data redif vealed ferences the presence of sporadic periods entirely void of points, and also to periods when points were between expected and empirical maximum distance ranges. It also explores the ability sca of ttthe ered sensors (i.e., hav in imeasuring ng varying d the istanc preceding e values trfoll uck owing no obvi in different driving ous trend) situations. . To understa Finally nd , these detectio suggestions for ns, v futur ideo e foot work age are w made. as code See d for Appendix LWRR and A.4 LBRR for complete . Filter 6 used data tables. these codes to include only data collected in LWRR periods, and to exclude data collected during LBRR 3.1. Impacts of Filtering periods. Cameras malfunctioned at times, during which LWRR and LBRR could not be coded. Impacts Associated r on the ad number ar data w ofedata re disc points arded, arensurin e detailed, g methodolo before discussing gical consthe istency. effects on recor In brief, ded metrics: filters relative 1 and 6 w velocity ere based , signal-to-noise on manual v ratio ideo and codes, while distance. filters 2 and 3 were Figure 6 shows the sizes of datasets as a function of the filtering steps. The datasets of based on recorded radar metrics, making them the most interesting ones in terms of radar trucks 2 and 3 were affected similarly. Downsampling (filter 5) included a mean of six data operation. Filters 4 and 5 were computational heuristics. See Appendix A.3 for more de- points and a mode of seven. Figure 7 is a distance histogram showing the effects of all tails. filtering steps. Note how early filters mostly remove measurements at close distances. 3. Results and Discussion This section explores effects of the filtering process, before discussing differences be- tween expected and empirical maximum distance ranges. It also explores the ability of the sensors in measuring the preceding truck in different driving situations. Finally, sugges- tions for future work are made. See Appendix A.4 for complete data tables. Appl. Syst. Innov. 2023, 6, x FOR PEER REVIEW 11 of 29 Appl. Syst. Innov. 2023, 6, x FOR PEER REVIEW 11 of 29 3.1. Impacts of Filtering 3.1. Impacts of Filtering Impacts on the number of data points are detailed, before discussing the effects on Impacts on the number of data points are detailed, before discussing the effects on recorded metrics: relative velocity, signal-to-noise ratio and distance. recorded metrics: relative velocity, signal-to-noise ratio and distance. Figure 6 shows the sizes of datasets as a function of the filtering steps. The datasets Figure 6 shows the sizes of datasets as a function of the filtering steps. The datasets of trucks 2 and 3 were affected similarly. Downsampling (filter 5) included a mean of six Appl. Syst. Innov. 2023, 6, 55 of trucks 2 and 3 were affected similarly. Downsampling (filter 5) included a mean of 11 of six 29 data points and a mode of seven. Figure 7 is a distance histogram showing the effects of data points and a mode of seven. Figure 7 is a distance histogram showing the effects of all filtering steps. Note how early filters mostly remove measurements at close distances. all filtering steps. Note how early filters mostly remove measurements at close distances. Figure 6. Number of measurements after each filtering step has been applied. Figure 6. Number of measurements after each filtering step has been applied. Figure 6. Number of measurements after each filtering step has been applied. Figure 7. Histogram of measured distances as a function of filtering steps. Figure 7. Histogram of measured distances as a function of filtering steps. Figure 7. Histogram of measured distances as a function of filtering steps. 3.1.1. Relative Velocity Overall, filtering reduced relative velocity data spread. Still, filter 1 increased relative velocity and data spread for trucks 2 and 3. In breaks, trucks were parked behind one another. Thus, removal of data from break periods serves to increase data spread, since these metrics differed more in driving periods. Excluding filter 1 (Driving and Following), filter 2 (relative velocity  30 km/h) affects the measured relative velocity the most, by Appl. Syst. Innov. 2023, 6, 55 12 of 29 removing data points corresponding to oncoming vehicles. This filter also reduced the average relative velocity to approximately 0.5 km/h for both trucks. Thus, while average relative velocities approach zero (as they should in car-following situations), they remain slightly negative. The negative value is due to most detected objects (with the exception of the preceding trucks) heading toward the sensors (as opposed to receding from them). The curated datasets for both trucks had slightly more instances of negative than positive relative velocities. The impact of filter 2 is greater for truck 3 than for truck 2, despite both having approximately equal sizes of datasets. It removes 20% of measurements from truck 3 and only 12% from truck 2, versus the number of measurements remaining after filter 1. In fact, the dataset of truck 3 enters filter 2 with a larger average relative velocity, presumably due to platoon instability and weight differences in which perturbations caused harsh braking for truck 3, which would naturally tend to occur at short distances, which filter 2 ended up removing (cf. Figure 7). Subsequent filtering steps slightly reduce the variability of relative velocity measurements, suggesting that erroneous detections are gradually removed. 3.1.2. Signal-to Noise Ratio (SNR) Filtering caused average SNR to stabilize around 29–32 dB for all trucks. Overall, filtering decreased SNR data spread. The minimum SNR was only affected by filter 3 (SNR < 15 dB). As intended, filter 3 subjected the data to a step-change, starting at 6.7–6.9 dB and ending up for all three trucks at 15.1 dB. The effects of each filter grad- ually diminish. Interestingly, all trucks measured different maximum SNR values, which were reduced in filter 5 (downsampling). As shown in Figure 5b, SNR filtering worked as intended by removing measurements at short distances. 3.1.3. Distance Distance was the only recorded metric which was not used as a basis for filtering. Average inter-vehicle distance values, shown in Table 3, suggest that truck 3 drove closer to its preceding truck than what truck 2 did. This was visually confirmed from video footage. After curation, average values were 38.6 and 36.1 m, for trucks 2 and 3, respectively. Still, distributions of distances appear to differ somewhat, with spikes at 0–10 m for truck 2, and 10–20 m for truck 3. Table 3. Distance metrics (in meters) after each filtering step. Truck Number Filtering 2 3 Step Standard Standard Average Maximum Average Maximum Deviation Deviation Raw 26.5 74.4 17.6 25.2 71.5 15.9 1 33.5 74.4 14.8 29.3 71.5 14.5 2 35.5 74.4 14.3 34.4 71.5 12.3 3 37.4 74.4 13.0 35.5 71.5 11.5 4 37.2 74.4 12.9 35.3 71.5 11.5 5 38.2 74.3 13.0 36.1 71.4 11.5 6 38.6 74.3 12.9 36.1 71.4 11.3 Unfiltered maximum distance values for trucks 2 and 3 were 74.4 and 71.5 m, respec- tively, and were virtually unaffected by filtering. Only filter 5 (downsampling) reduced maximum distances, and only by 0.1 m in both cases. This resulted in curated maximum distances of 74.3 and 71.4 m. While distance ranges were never systematically tested, maxi- mum distance values may be used as a proxy. While driving on straight road segments, inter-vehicle distances oscillated. The trucks would occasionally drive closely together before becoming dispersed, travelling with spacing between the trucks so large that each preceding truck was eventually located beyond sensor range. Thus, for each truck, measure- Appl. Syst. Innov. 2023, 6, 55 13 of 29 ments should exist at the radar range boundary. As stated, filter 6 (LWRR) did not further reduce maximum distance values than what filter 5 did. Figure 7 shows that filter 6 did not further reduce maximum distance values since it had little effect beyond approximately 50–55 m. Since LWRR video coding was carried out visually, somewhat imprecisely, the farthest radar measurements still tend to appear in LWRR-filtered data. Had it been more precise, maximum distance values following filter 5 would best represent actual upper sensor ranges. The aforementioned maximum distance values for trucks 2 and 3 fall short of the theoretical ranges 75.0 and 73.1 m by 0.9% and 2.4%, respectively. Hence, both sensors appear to underperform slightly versus expected ranges, but still fall within the estimate provided by the supplier. The deviation is smallest for truck 2. Since the datasets from both trucks are otherwise comparable, the chosen parameter values N and f for the radar s 0 sensor in truck 2 appears to be preferable. Future testing could explore this. For the curated radar data, average relative velocity and average SNR were inspected as a function of inter-vehicle distances. For both trucks 2 and 3, these metrics were calcu- lated within successive 10-m distance bins. The lower bin was 0–10 m, and the upper bin was 70 m and above. For all bins, average relative velocities fell within the error margin of the sensors. On the other hand, average SNR values are more interesting. For all bins, average SNR very seldomly fell below 30 dB, and remained high even at long distances, indicating that the radar cross-sections of the preceding trucks are sufficiently large to allow for longer detection ranges, cf. Tables A8 and A9 in Appendix A.4. Thus, parameters could likely have been chosen to achieve a maximum range approaching, and perhaps even exceeding, 100 m. The radar user manual illustrates that cars can be detected at 75 m, while buildings can be detected at 100 m. Due to the size of the truck rear walls, they may provide ‘building-like’ detection ranges. This study tracked large, reflective metal back walls of semi-trailers, with cross-sectional areas exceeding 8 square meters. The back walls of semi-trailers are larger than those of other vehicles and presumably facilitate higher-quality detections. Measuring distances to smaller vehicles should be tested. While the maximum distance range may be greater when detecting trucks than passenger cars, situations may exist where trucks are less favorable. For instance, when traversing sharp curves at short distances, the radar cross-section changes as truck back walls change angles in relation to the sensors, as opposed to being located perpendicular to them. Some military ships and aircraft are deliberately made from planar surfaces joined at sharp angles to achieve radar stealth. Similarly, truck back walls are also two-dimensional planar surfaces. In curves, reflected signals may scatter away from the receivers and cause data loss. This may also occur if the preceding truck is located outside the main lobe of the radar antenna. Such diffraction effects were reported in [38]. 3.2. Radar Sensor Operation in Different Driving Situations The curated radar data were coupled with video codes to determine whether the sensors captured high-quality inter-vehicle distance measurements when they should have been able to. Filter 6 aimed to remove periods when respective preceding trucks were located beyond radar range. We now check whether the radars were able to account for the remaining duration. Driving is explored, and also Tunnel and Roundabout subcategories. Outputs from filter 6 contain radar data at 1 Hz, so aggregated video code durations (in seconds), when filtered for LRWW, are directly comparable to curated radar data. The proportions are shown in Table 4. Both sensors retained a similar number of measurements in each condition. As Table 4 shows, the sensors detected the preceding truck in most situations where it had the opportunity to do so. Using Tunnels to illustrate: Trucks 2 and 3 drove in Tunnels while having their respective preceding trucks within radar range (LWRR) for approximately 26 min (1647 and 1688 s, respectively). Aggregated over these same periods, their radar sensors outputted 1576 and 1609 curated measurements (at 1 Hz) after filter 6. As proportions, this yields 96% and 95%, respectively. Examples from each condition are shown in Table A5 in Appendix A.2. Appl. Syst. Innov. 2023, 6, 55 14 of 29 Table 4. Proportions of curated radar data retained vs. LWRR-filtered aggregated event durations. Trucks Condition 2 3 Driving * 85% 83% Tunnels 96% 95% Roundabouts 88% 89% * Driving includes Tunnels and Roundabouts. Radar data and video codes can be used to visualize and explore excerpts of curated radar data (LWRR) from different driving intervals. Inter-vehicle distances are plotted versus time. Excerpts indicate large variability in inter-vehicle distances as a function of infrastructure type and road standard, as suggested in [23]. Since Tunnel and Roundabout video codes have non-zero durations, a choice had to be made regarding from which truck to visualize them. Truck 2 was chosen, being the middle truck in the platoon for 88% of the drive. Note that, at the time resolution used (1 min divisions), the difference would have been negligible if visualizing codes from one of the other two vehicles. Oncoming trucks seemed to cause the platoon to slow down, which also affected inter-vehicle distances, particularly on narrow roads and in sharp curves. Therefore, Oncoming truck video codes are shown. Being point events, these could be visualized for each truck separately. The horizontal axes (d hh:mm) were not fixed, so excerpts have slightly different durations (between 5 to 10 min). As the field study took place over two days, d-values are either 1 or 2. 3.2.1. Visual Verification of Maximum Range Figure 8 shows three excerpts from Driving. While no tunnels or roundabouts are shown, statistics from this category in Table 4 also include those durations, thus showcasing diverse, complex driving segments. While serving as followers for 7 h, trucks 2 and 3 drove with their respective preceding trucks within radar range for approximately 6 h. All excerpts in Figure 8 had the preceding truck within range, so trends for measured distances appear mostly continuous. The top excerpt stems from an old, narrow road section without centerlines. Trucks often adjusted their speeds, including when encountering opposing trucks, as revealed by reduced inter-vehicle distances. The middle excerpt stems from the traversal of a flat and wide high-quality road with a 90 km/h speed limit and gentle horizontal curves. Here, the trucks maintained constant distances over long time periods, and opposing trucks did not influence the platoon. The individual data points from truck 2 which are located below the general trend were chosen by the algorithm since the sensor did not detect other data points during that logging instance. In general, truck 2 had more such erroneous measurements than truck 3 did, but it is unclear why this is the case. The lower excerpt illustrates the descent of a challenging mountain pass (negative 6% gradient), where trucks 2 and 3 reduced their speed repeatedly, to avoid speeding and becoming located too close to their respective preceding truck. These excerpts suggest that truck platooning is more suitable on wide, modern roads than on old roads with adverse horizontal and vertical alignment. Appl. Syst. Innov. 2023, 6, x FOR PEER REVIEW 15 of 29 Appl. Syst. Innov. 2023, 6, 55 15 of 29 that truck platooning is more suitable on wide, modern roads than on old roads with ad- verse horizontal and vertical alignment. Figure 8. Inter-vehicle distances measured from three separate driving excerpts, LWRR, during Figure 8. Inter-vehicle distances measured from three separate driving excerpts, LWRR, during which which radar range was never exceeded. radar range was never exceeded. Figure 9 shows three excerpts during which, for both trucks 2 and 3, the inter-vehicle Figure 9 shows three excerpts during which, for both trucks 2 and 3, the inter-vehicle distance appears to have occasionally exceeded maximum radar ranges. Blue (truck 2) and distance appears to have occasionally exceeded maximum radar ranges. Blue (truck 2) green (truck 3) horizontal lines illustrate the farthest distances detected, which differ and green (truck 3) horizontal lines illustrate the farthest distances detected, which differ somewhat somewhat between between exce excerpts. rpts. Nonetheless, Nonetheless, for forradar radarsensors sensorsin intr tr ucks ucks2 2and and3, 3, re respectively spectively, , the the m maximum aximum distances distances ar eare appr ap oximately proximately 75 and 75 and 70 m, 70 m, wh which support ich support the the aforementioned aforemen- statistics-based tioned statistics-based r radar range adestimates. ar range estim The thr ates. The three excerpts stem f ee excerpts stem from two dif rom two di ferent mountain fferent passes. mounta The in passes. The two upper excerpts corre two upper excerpts correspond to traversal spond to tra of Mountain versal ofpass Mount A, a with in pa very ss A, dif with v ficultecombinations ry difficult com ofbsharp inations o horizontal f sharp horizont curves and al c vertical urves and v gradient ertical s. V gr ideos adient wer s. V e ialso deos useful in inspecting the radar data after curation. The combination of long inter-vehicle were also useful in inspecting the radar data after curation. The combination of long inter- distances and horizontal curves occasionally caused the preceding truck to be obscured by vehicle distances and horizontal curves occasionally caused the preceding truck to be ob- rock walls at the inner part of right-turn curves (pale shading). Mountain pass B, shown in scured by rock walls at the inner part of right-turn curves (pale shading). Mountain pass the lowermost excerpt, was more forgiving in terms of road alignment. B, shown in the lowermost excerpt, was more forgiving in terms of road alignment. In the upper excerpt, the data from truck 3 (green) at timestamp 16:56 potentially reveals the presence of a phenomenon which, together with the high average SNR values measured at large distances, suggests that sensors are capable of measuring the preceding truck far away. It appears as if measurements which naturally belong to the top of the green curve are folded down, instead of occurring at 90–100 m, where extrapolation would place them. Thus, it looks like the radar in truck 3 does measure the preceding truck, despite it being located beyond the maximum range imposed by the chosen parameters. Stated otherwise, the radar appears to measure points beyond its unambiguous range, referring to the maximum distance a target can have while it can be guaranteed that the reflected pulse from that target corresponds to the most recent transmitted pulse [14,39]. At this timestamp, it appears as if the returned signal is associated with the wrong transmitted pulse, so Appl. Syst. Innov. 2023, 6, 55 16 of 29 the range becomes ambiguous. Filtering for LWRR and LBRR served to remove most durations where such long-distance samples may have been folded into the ambiguous Appl. Syst. Innov. 2023, 6, x FOR PEER REVIEW 16 of 29 range. However, the manual video coding process did not remove all such instances. The presence of folding may have influenced distance metrics to the downside. Figure 9. Inter-vehicle distances measured from three separate driving excerpts, LWRR, during Figure 9. Inter-vehicle distances measured from three separate driving excerpts, LWRR, during which which radar range was exceeded. Horizontal lines show farthest measurements. radar range was exceeded. Horizontal lines show farthest measurements. In the upper Data from truck excerpt, the 2 in the da middle ta from excerpt truck 3 (green) of Figure 9 atar tiemest noisier amp 16 than :56 potentia for all other lly reveals the presence of a phenomenon which, together with the high average SNR values excerpts shown, but it is unclear whether the folding phenomenon occurs here. The outage measured at large distances, suggests that sensors are capable of measuring the preceding for truck 2 in the lower excerpt stems from a period when a passenger car partly overtook truck the platoon, far away. It being a sandwiched ppears as ifbetween measurements whi truck 1 andc2h until natura also lly belong to the top of overtaking the lead tr the uck. green c Data from urvsuch e areperiods folded d wer own, in e removed steadby of occur filter 1 r (Driving ing at 90–100 m and Following , where ). extrapolation would place them. Thus, it looks like the radar in truck 3 does measure the preceding 3.2.2. Tunnels truck, despite it being located beyond the maximum range imposed by the chosen param- eters. St It isated unclear otherwise, th how tunnels e radaf ar fect appears to measure points bey the ability of the sensors to o measur nd its un e inter ambiguo -vehicle us distances. Tunnels may reduce operational complexity, as rock walls cause peripheral range, referring to the maximum distance a target can have while it can be guaranteed narrowing of roadside areas. However, walls, lighting and ventilation elements may that the reflected pulse from that target corresponds to the most recent transmitt ed pulse introduce clutter. Such features are less frequent on roads in natural terrain. [14,39]. At this timestamp, it appears as if the returned signal is associated with the wrong All Tunnel driving occurred in truck order 1–2–3. Figure 10 shows excerpts from transmitt ed pulse, so the range becomes ambiguous. Filtering for LWRR and LBRR served six representative Tunnel traversals (green shading). Driving periods outside tunnels have to remove most durations where such long-distance samples may have been folded into white backdrops. All excerpts were coded as LWRR, except for the period between the the ambiguous range. However, the manual video coding process did not remove all such two tunnels in the top excerpt, when data were lacking for truck 2. This period was instances. The presence of folding may have influenced distance metrics to the downside. coded as LBRR, as truck 1 was located far away. Comparing Figures 8 and 10, it seems Data from truck 2 in the middle excerpt of Figure 9 are noisier than for all other ex- as if filter 4 (target selection) chooses erroneous data points at comparable frequencies cerpts shown, but it is unclear whether the folding phenomenon occurs here. The outage both inside and outside tunnels. Thus, tunnel driving does not appear to degrade radar for truck 2 in the lower excerpt stems from a period when a passenger car partly overtook operating conditions. In tunnels, maximum distance values for trucks 2 and 3 were reduced the platoon, being sandwiched between truck 1 and 2 until also overtaking the lead truck. by 8–9%, and the distance standard deviation dropped by 18% and 15%, respectively. Data from such periods were removed by filter 1 (Driving and Following). Thus, inter-vehicle distances were moderated by Tunnels, causing closer, more uniform driving at lower speeds. This made preceding trucks occupy a larger part of radar the field- 3.2.2. Tunnels It is unclear how tunnels affect the ability of the sensors to measure inter-vehicle dis- tances. Tunnels may reduce operational complexity, as rock walls cause peripheral nar- rowing of roadside areas. However, walls, lighting and ventilation elements may intro- duce clutt er. Such features are less frequent on roads in natural terrain. Appl. Syst. Innov. 2023, 6, x FOR PEER REVIEW 17 of 29 All Tunnel driving occurred in truck order 1–2–3. Figure 10 shows excerpts from six representative Tunnel traversals (green shading). Driving periods outside tunnels have white backdrops. All excerpts were coded as LWRR, except for the period between the two tunnels in the top excerpt, when data were lacking for truck 2. This period was coded as LBRR, as truck 1 was located far away. Comparing Figures 10 and 8, it seems as if filter 4 (target selection) chooses erroneous data points at comparable frequencies both inside and outside tunnels. Thus, tunnel driving does not appear to degrade radar operating condi- tions. In tunnels, maximum distance values for trucks 2 and 3 were reduced by 8–9%, and the distance standard deviation dropped by 18% and 15%, respectively. Thus, inter-vehi- Appl. Syst. Innov. 2023, 6, 55 17 of 29 cle distances were moderated by Tunnels, causing closer, more uniform driving at lower speeds. This made preceding trucks occupy a larger part of radar the field-of-view. How- ever, measurements collected in tunnels were noisier than those collected during Driving, of-view. However, measurements collected in tunnels were noisier than those collected indicated by lower SNR values (weaker signal). For both followers, Tunnel filtering re- during Driving, indicated by lower SNR values (weaker signal). For both followers, Tunnel duced mean SNR by 3%, maximum values by 6–8% and standard deviations by 7–11%. filtering reduced mean SNR by 3%, maximum values by 6–8% and standard deviations by This did not affect the curated radar data when plott ed: In Figure 9, erroneously selected 7–11%. This did not affect the curated radar data when plotted: In Figure 9, erroneously targets in tunnels appear to have distance values at similar deviations to the trend, com- selected targets in tunnels appear to have distance values at similar deviations to the trend, pared to erroneously selected targets outside tunnels. Thus, it appears as if inter-vehicle compared to erroneously selected targets outside tunnels. Thus, it appears as if inter-vehicle distance measurements between platooning trucks are not adversely affected by tunnels. distance measurements between platooning trucks are not adversely affected by tunnels. Figure 10. Radar data for three separate truck driving excerpts, Tunnel and LWRR. Figure 10. Radar data for three separate truck driving excerpts, Tunnel and LWRR. 3. 3.2.3. 2.3. Ro Roundabouts undabouts Roundabouts may allow for exploring the effects of sharp road curvature on radar Roundabouts may allow for exploring the effects of sharp road curvature on radar operation. Most roundabout traversals involved straight movements, encountering little to operation. Most roundabout traversals involved straight movements, encountering litt le no traffic. Figure 11 shows excerpts from five traversals (pale red shading), all of which to no traffic. Figure 11 shows excerpts from five traversals (pale red shading), all of which occurred in truck order 1–2–3. Driving periods outside roundabouts have white backdrops. occurred in truck order 1–2–3. Driving periods outside roundabouts have white back- The first traversal in the middle excerpt involved trucks 1 and 2 performing a full revolution drops. The first traversal in the middle excerpt involved trucks 1 and 2 performing a full to get rid of external vehicles located between trucks 2 and 3. Truck 3 had Other vehicles revolution to get rid of external vehicles located between trucks 2 and 3. Truck 3 had Other preceding it before entering this roundabout, so its data were removed by filter 1. The vehicles preceding it before entering this roundabout, so its data were removed by filter 1. traversal in the lower excerpt involved all three trucks revolving one round. Trucks 2 and The traversal in the lower excerpt involved all three trucks revolving one round. Trucks 2 3 both had 10 traversals as followers, with 5 straight, 1 left turn and 4 right turns. For and 3 both had 10 traversals as followers, with 5 straight, 1 left turn and 4 right turns. For both trucks, only half of the aggregated Roundabout durations had the preceding truck within field-of-view, and even when accounting for field-of-view, 11–12% of data are lost. Figure 11 shows that data points retained in roundabouts generally have scattered distance values which are too large to represent the preceding truck. Sharp curve radii and limited antenna beam width resulted in lost field-of-view to the preceding truck, so radar sensors detected irrelevant objects until field-of-view was regained after the turn. Appl. Syst. Innov. 2023, 6, x FOR PEER REVIEW 18 of 29 both trucks, only half of the aggregated Roundabout durations had the preceding truck within field-of-view, and even when accounting for field-of-view, 11–12% of data are lost. Figure 11 shows that data points retained in roundabouts generally have scatt ered dis- tance values which are too large to represent the preceding truck. Sharp curve radii and Appl. Syst. Innov. 2023, 6, 55 18 of 29 limited antenna beam width resulted in lost field-of-view to the preceding truck, so radar sensors detected irrelevant objects until field-of-view was regained after the turn. Figure 11. Radar data from three separate driving excerpts with Roundabouts, LWRR. Figure 11. Radar data from three separate driving excerpts with Roundabouts, LWRR. Average SNR values were 30 and 32% lower for trucks 2 and 3 in Roundabouts than Average SNR values were 30 and 32% lower for trucks 2 and 3 in Roundabouts than for for Driving, respectively, so the radar sensors detected noisier data. Proportions of radar Driving, respectively, so the radar sensors detected noisier data. Proportions of radar data data as a function of LWRR-filtered Roundabout durations (88–89%), are greater than for as a function of LWRR-filtered Roundabout durations (88–89%), are greater than for Driving Driving (83–85%). This is likely due to the trucks being grouped closer together when trav- (83–85%). This is likely due to the trucks being grouped closer together when traversing ersing Roundabouts. Speeds were also lower, causing smaller inter-vehicle distances and Roundabouts. Speeds were also lower, causing smaller inter-vehicle distances and greater greater spatial concentration of measurements. For trucks 2 and 3, mean distances were spatial concentration of measurements. For trucks 2 and 3, mean distances were 66–62% 66–62% shorter during Roundabouts than during Driving. The exploration shows that pre- shorter during Roundabouts than during Driving. The exploration shows that preceding ceding trucks are tracked poorly when the platoon passes through roundabouts. trucks are tracked poorly when the platoon passes through roundabouts. 3.3. Suggestions for Future Work 3.3. Suggestions for Future Work This study explored the extent to which uRAD radar sensors could capture inter- This study explored the extent to which uRAD radar sensors could capture inter- vehicle distances during truck platooning. Several suggestions have been identified. vehicle distances during truck platooning. Several suggestions have been identified. First, mode 4 (7 Hz) might be more suitable than mode 3 (13 Hz). As data were First, mode 4 (7 Hz) might be more suitable than mode 3 (13 Hz). As data were downsampled anyway, the trade-off between higher update rate (mode 3) and enhanced downsampled anyway, the trade-off between higher update rate (mode 3) and enhanced properties for ghost target reductions in complex scenarios (mode 4) should be explored. properties for ghost target reductions in complex scenarios (mode 4) should be explored. To simplify data collection, radars could be remotely engaged from escort vehicles. If To simplify data collection, radars could be remotely engaged from escort vehicles. field trials are undertaken in areas with adequate cell coverage, virtual network computing If field trials are undertaken in areas with adequate cell coverage, virtual network compu- (VNC) could be used to remote control all Raspberry Pi microprocessors. Radar data may ting (VNC) could be used to remote control all Raspberry Pi microprocessors. Radar data also be visualized and coded in real-time. Future work may validate the radar data against may also be visualized and coded in real-time. Future work may validate the radar data GNSS positions, if undertaken in areas with good conditions for GNSS receivers. The against GNSS positions, if undertaken in areas with good conditions for GNSS receivers. performance of the sensors should also be compared against a known baseline, i.e., radars The performance of the sensors should also be compared against a known baseline, i.e., with known characteristics, such as those listed in [14], in controlled environments. radars with known characteristics, such as those listed in [14], in controlled environments. The cross-sectional signature of the trucks was not measured, and scattering effects were not explored. If present, such effects would be reflected in SNR values if systematically inspected at equal distances while varying the angle of the back wall of the preceding truck, resembling the set-up in [40]. Windshield attenuation effects were also not studied. Synchronization of videos and radar data, and the subsequent process of video coding, worked well. However, both should preferably be automated, to reduce post-processing Appl. Syst. Innov. 2023, 6, 55 19 of 29 efforts and related human errors. Traffic and infrastructure events may be identified directly from radar data. Herein, tunnels and roundabouts were coded from the moment where each respective truck entered them, and to the moment when each respective truck left them. Later, video codes were overlaid on the radar data collected from each truck. However, since radar data shows the preceding truck, this may have introduced a systematic error. Perhaps video codes from the preceding truck should have been used instead. For tunnels, this is not particularly problematic since tunnel traversals (with LWRR) had long durations (on average 1.1 min). Thus, entering and leaving the tunnel occupies a very small part of the 1 1 total duration. For roundabouts, however, the preceding truck had often traversed – of 4 2 the roundabout before the truck in question entered it, and it was first coded. Distinctions may be made between the preceding truck being located beyond range, and it being located laterally beyond field-of-view, as these are different phenomena. All data points may also be given metadata for all relevant video codes, simplifying video inspections of interesting events in the data. Statistical approaches may allow cut-off values for filters 2 and 3 to be chosen auto- matically. In the target detection step, distance filtering could also be considered, perhaps discarding data points with distances deviating significantly from the general trend. This may solve the problem of erroneous single-object detections. Established data filtering, target tracking [41] or clustering techniques [42–44] could also be used, alongside more computationally complex methods for annotating or labelling combinations of radar and camera footage, e.g., in [31,45], and perhaps also machine learning approaches [46]. At times, dashboard cameras malfunctioned due to power issues, totaling 6% of driving time, during which 10% of all radar data were logged. As it was not possible to de- termine whether the sensors had reasonable operating conditions in these periods, the data were discarded. Mitigations include redundant cameras and independent power supplies. 4. Conclusions Anteral uRAD radar sensors for Raspberry Pi were tested for estimating inter-vehicle distances between trucks. Three trucks participated in a real-word platooning field study. Data from integrated sensors were unavailable. Comparable results were found from the sensors in the two rearmost trucks, suggesting that they are feasible for this use case. Data filtering involved a multi-faceted methodology. While also filtering based on relative velocity and signal-to-noise ratio, video footage allowed for removal of data from irrelevant periods, and for exploring sensor operation in roundabouts and tunnels. This would not have been possible without video footage. The curated radar data can be used to model expected fuel savings from truck platooning on specific types of roads and road features. Sensor ranges were estimated at 74 and 71 m, i.e., slightly shorter than suggested by theoretical calculations. The sensors captured the preceding truck for 83–85% of the time when it was located within radar range. In tunnels specifically, 95–96% of driving time was accounted for, likely due to closer driving. Average SNR decreased 3% in tunnels, compared to all driving, but this did not appear to affect the accuracy of the target detection step. When turning in roundabouts, the field-of-view to the preceding truck was often lost, and the sensors detected their surroundings until field-of-view was regained after completing the turn, causing average SNR values to drop (30–32% lower). Straight movements in roundabouts were less problematic, as field-of-view was mostly retained. High SNR values were observed at far distances, indicating that the sensors, with optimal parameters, may be capable of measuring preceding trucks further away. The findings suggest that simple, inexpensive radar sensors and action cameras can facilitate collection of inter-vehicle distance data from truck platooning field trials. Appl. Syst. Innov. 2023, 6, x FOR PEER REVIEW 20 of 29 findings suggest that simple, inexpensive radar sensors and action cameras can facilitate collection of inter-vehicle distance data from truck platooning field trials. Appl. Syst. Innov. 2023, 6, 55 20 of 29 Author Contributions: Conceptualization, M.M.L.; methodology, M.M.L., T.T. (Thomas Thoresen), M.H.R.E.; software, T.T. (Thomas Thoresen); validation, T.T. (Thomas Thoresen); formal analysis, M.M.L.; investigation, M.M.L.; data curation, T.T. (Thomas Thoresen); writing—original draft prep- aration, M.M.L.; writing—review and editing, M.M.L., M.H.R.E., T.T. (Trude Tørset), T.L.; visuali- Author Contributions: Conceptualization, M.M.L.; methodology, M.M.L., T.T. (Thomas Thoresen), zation, T.T. (Thomas Thoresen); supervision, T.T. (Trude Tørset), T.L.; project management, M.M.L.; M.H.R.E.; software, T.T. (Thomas Thoresen); validation, T.T. (Thomas Thoresen); formal analysis, funding acquisition, T.L. All authors have read and agreed to the published version of the manu- M.M.L.; investigation, M.M.L.; data curation, T.T. (Thomas Thoresen); writing—original draft prepa- script. ration, M.M.L.; writing—review and editing, M.M.L., M.H.R.E., T.T. (Trude Tørset), T.L.; visualization, T.T. (Thomas Thoresen); supervision, T.T. (Trude Tørset), T.L.; project management, M.M.L.; funding Funding: This research was funded in part by the Norwegian University of Science and Technology acquisition, T.L. All authors have read and agreed to the published version of the manuscript. (NTNU) and the Norwegian Public Roads Administration (NPRA), through the Innovation and Im- plementation initiative (D11351). The field study itself, titled “Smart and Connected Truck Train” Funding: This research was funded in part by the Norwegian University of Science and Technology was financed by the NPRA, cf. the Norwegian database for public procurement. Notice of intent for (NTNU) and the Norwegian Public Roads Administration (NPRA), through the Innovation and procurement of transportation research services. “Smart and Connected Truck Train”. Norwegian Implementation initiative (D11351). The field study itself, titled “Smart and Connected Truck Train” national database for public procurement: htt ps://doffin.no/nn/Notice/Details/2020-301190 (ac- was financed by the NPRA, cf. the Norwegian database for public procurement: https://doffin.no/ cessed 3 May 2023). nn/Notice/Details/2020-301190. Institutional Review Board Statement: Not applicable. Institutional Review Board Statement: Not applicable. Informed Consent Statement: Informed consent was obtained from all participants involved. Informed Consent Statement: Informed consent was obtained from all participants involved. Data Availability Statement: Data are available on request from the corresponding author. Data Availability Statement: Data are available on request from the corresponding author. Acknowledgments: The main author would like to acknowledge the Anteral supplier team for an- Acknowledgments: The main author would like to acknowledge the Anteral supplier team for swering inquiries. answering inquiries. Conflicts of Interest: The authors declare no conflict of interest. Conflicts of Interest: The authors declare no conflict of interest. Appendix A Appendix A The appendix is subdivided into four parts, providing background information and The appendix is subdivided into four parts, providing background information and methodological details for reproducibility, alongside tables with results. methodological details for reproducibility, alongside tables with results. Appendix A.1. The Radar Sensors Appendix A.1. The Radar Sensors Anteral uRAD radar sensors, version 1.1, were used alongside Raspberry Pi 4, model Anteral uRAD radar sensors, version 1.1, were used alongside Raspberry Pi 4, model B B with 4 GB RAM. Micro-SD cards (16 GB) were purchased, with the Raspberry Pi oper- with 4 GB RAM. Micro-SD cards (16 GB) were purchased, with the Raspberry Pi operating ating system and uRAD software pre-installed. Technical support was also purchased. Ta- system and uRAD software pre-installed. Technical support was also purchased. Table A1 ble A1 provides an overview of the radar configuration parameters. Testing prior to the provides an overview of the radar configuration parameters. Testing prior to the field trial is field trial is outlined in Figure A1 and Table A2. Collected test data and corresponding outlined in Figure A1 and Table A2. Collected test data and corresponding camera footage camera footage were sent to the supplier, who proposed recommendations for future test- were sent to the supplier, who proposed recommendations for future testing. Table A3 ing. Table A3 details the parameters which were used in the field study. details the parameters which were used in the field study. (a) (b) (c) Figure A1. (a) Dump truck test set-up; (b) Back wall of preceding truck seen from cabin during Figure A1. (a) Dump truck test set-up; (b) Back wall of preceding truck seen from cabin during driving, with radar sensor (yellow) and dashcam (red) circled; (c) Radar sensor side view. driving, with radar sensor (yellow) and dashcam (red) circled; (c) Radar sensor side view. The supplier detailed the pre-processing steps taken before sensors store data to mem- ory: The radar transceiver chip receives the reflected signal. The mixer mixes the received (RX) signal with the transmitted (TX) signal, and outputs in-phase (I) and quadrature (Q) components at Intermediate Frequency (IF). These two analog I/Q IF signals go through a low-pass filter, an amplifier stage and a high-pass filter. Filter values are proprietary. Subse- quently, the analog signal is digitalized with an ADC at 25 kHz in mode 1 and 200 kHz in Appl. Syst. Innov. 2023, 6, 55 21 of 29 modes 2, 3 and 4. In the digital domain, the complex signal is formed, FFT obtained, and from it, range, velocity and SNR of detected targets are derived. The supplier also stated that calibration by the user is not needed, as performance is controlled in the lab prior to shipment, by measuring a constant distance of 1.5 m in mode 2. The combination of one Raspberry Pi attached to one uRAD radar board is here termed a device. The device was fixed to the interior of the windshield using a universal phone suction mount with a flexible arm, oriented such that the USB-C and micro-HDMI ports pointed directly upwards (cf. Figure A1). Since Raspberry Pi microprocessors were powered on and off using a USB-C cable, this orientation facilitated easy access and line of sight from above, for cable insertion and removal. It also made sure that the sensor did not detect the mount itself, being located outside radar field-of-view. Powerbanks and USB-C cables powered the devices at optimum voltage and amperage, while making data logging independent of truck power systems and status. This eliminated potential issues with undervolting and voltage spikes from in-vehicle outlets, while leaving flexibility for when logging start and stop had to be administered. Each Raspberry Pi had a Bluetooth USB dongle for a wireless keyboard and mouse, minimizing direct device contact. The dongle added an additional reason for ensuring stable power supply. The radar supplier confirmed that the dongle radio frequency would not affect radar operation. Along with the dongle, the micro-HDMI and power cables remained plugged into the device throughout the field trial. This allowed the devices to remain vehicle-mounted throughout both days, requiring only insertion and removal of the far-ends of the cables into the screen and powerbank for interfacing with the devices and powering them devices on and off, respectively. Radar output files never exceeded 5 MB, i.e., they were unproblematic with respect to SD card storage capacity. A battery powered portable monitor was used. With 30 vertical fields-of-view, 15 swept down from horizontal, so vertical sweep became 75 , assuming sensors were mounted perfectly level. If trucks, with uRAD antennas at height h , traversed a constant gradient road section, the road would be detected at a distance, d, given Equation (A1): d = h tan(75 ) (A1) The presence of aftermarket dashboard tabletop surfaces required placing sensors at slightly different heights in each truck during the field study. With radar antennas at heights of 2.29 m (truck 2) and 2.15 m (truck 3), the road would be detected 8–8.5 m forwards. Having 30 fields-of-view also in the horizontal direction, radars also saw this far sideways at road level. Table A1. Overview and discussion of uRAD radar parameters. Parameter Discussion Of four modes available, only modes 3 (triangular) and 4 (dual-rate) measured both distance and velocity. Velocity would enable filtering away stationary and oncoming objects, to be left with desired inter-vehicle distance to the preceding truck. Modes 3 and 4 differed in upper distance range and update rate. Mode 3 had an upper distance range of 100 m, versus 75 m for mode 4. The supplier stated Mode that the range would also depend on the target, meaning its radar cross-section: “(. . . ) a person is detected up to 40 m. (. . . ) a truck, that is bigger and reflects more, (. . . ) will be detected [at] 70 m but probably (. . . ) much farther. 100 m is not a limitation of the radar, [but] a guide (. . . ) for very big targets.” Mode 4 should reduce ghost target detections in multi-target scenarios, at the expense of reduced range. Appl. Syst. Innov. 2023, 6, 55 22 of 29 Appl. Syst. Innov. 2023, 6, x FOR PEER REVIEW 22 of 29 Table A1. Cont. Parameter Discussion Ramp start frequency, f , could be set as 5–195 for modes 2–4. Operation Ramp start frequency, f0, could be set as 5–195 for modes 2–4. Operation bandwidth, BW, meaning bandwidth, BW, meaning the frequency sweep used in modes 2–4, depends on f , the frequency sweep used in modes 2–4, depends on f0, and should be maximized, subject to and should be maximized, subject to Equation (A2), to increase accuracy and to Equation (A2), to increase accuracy and to distinguish closely located targets. For each radar, dif- distinguish closely located targets. For each radar, different values were chosen to Ramp start freq., f0. Ramp start ferent v freq., f a . lues were chosen to avoid interference. avoid interference. Operation band- Operation bandwidth, BW 𝐵𝑊 = 245𝑓 BW = 245 f (A2) (A2) max width, BW The f0 parameter denotes the starting frequency of the waves emitt ed by the sensor. The sensor op- The f parameter denotes the starting frequency of the waves emitted by the sensor. The sensor operates at a frequency bandwidth of 24.005–24.245 GHz, and erates at a frequency bandwidth of 24.005–24.245 GHz, and f0 values are defined (in MHz) to set f values are defined (in MHz) to set the offset from the lower threshold. the offset from the lower threshold N is the number of samples taken from the reflected wave to calculate distance Ns is the number of samp s les taken from the reflected wave to calculate distance and velocity. and velocity. Highest update rate requires lowest possible N . However, a trade-off Highest update rate requires lowest possible Ns. However, a trade-off is needed, since BW and Ns is needed, since BW and N determine maximum range, through Equation (A3). determine maximum range, through Equation (A3). Samples and ramp Ns 𝑐𝑒𝐷𝑖𝑠𝑡𝑎𝑛 =75∙ (A3) Samples and ramp duration, N s Distance = 75 (A3) max duration, Ns BW The Ns parameter serves two purposes. Firstly, it defines the duration of each wave ramp, and sec- The N parameter serves two purposes. Firstly, it defines the duration of each ondly, it outlines the sampling rate from the reflected wave, per ramp duration, which can be wave ramp, and secondly, it outlines the sampling rate from the reflected wave, used to calculate output metrics. per ramp duration, which can be used to calculate output metrics. Ntar is the number of targets that the sensor detects, 5 being maximum. If detecting more objects, Max. detected tar- N is the number of targets that the sensor detects, 5 being maximum. If detecting tar the sensor logs data for those 5 with highest SNR. Ntar was maximized, capturing most data and more objects, the sensor logs data for those 5 with highest SNR. N was tar gets, Ntar Max. detected targets, N tar providing possibility for filtering unwanted objects later. maximized, capturing most data and providing possibility for filtering unwanted objects later. For modes 2–4, Rmax is the maximum distance below which targets will be detected. Rmax artificially reduces the zone of interest, excluding For modes 2–4, R targets beyo is the maximum nd this distance distanc below e, evwhich en if they targets hav will e hbe igher SNR max detected. R artificially reduces the zone of interest, excluding targets beyond than those within it. Rmax was cho max sen as 100 for all sensors, as this would search targets within the this distance, even if they have higher SNR than those within it. R was chosen max Maximum detec- entire range. When asked if the sensors would stay fixed on the preceding truck in horizontal as 100 for all sensors, as this would search targets within the entire range. When tion distance, Rmax curves, the supplier stated that manual antenna modification could double the horizontal FOV, to asked if the sensors would stay fixed on the preceding truck in horizontal curves, Maximum detection distance, R max the detriment of upper detection range. No manual modifications were made. For vertical curves, the supplier stated that manual antenna modification could double the horizontal FOV, to the detriment of upper detection range. No manual modifications were the supplier cited that the road in front of the truck, which would be more visible in vertical sag made. For vertical curves, the supplier cited that the road in front of the truck, curves, could reflect the signal, masking the preceding truck. which would be more visible in vertical sag curves, could reflect the signal, Moving target indi- masking the preceding truck. cator, MTI. Moving target indicator (MTI) allowed for including data only from objects with motion relative Moving target indicator (MTI) allowed for including data only from objects with Moving target indi-cator, MTI. Movement detec- to the sensor. Mth is only relevant when using uRAD as a movement detector, and was not used. motion relative to the sensor. Mth is only relevant when using uRAD as a Movement detection, Mth tion, Mth movement detector, and was not used. Table A2. Pre-trial testing of radar parameter configurations. Table A2. Pre-trial testing of radar parameter configurations. Supplier Modifications and Recommen- Pre-Test Steps Parameters User Experience Supplier Modifications and dations Pre-Test Steps Parameters User Experience Recommendations • Switch off Mth, as it is not relevant for the application. Switch off Mth, as it is not • Preceding traffic recorded well. relevant for the application. • Replaced GUI with Python script for Preceding traffic recorded well. 1: Passenger car Mode = 2 Replaced GUI with Python script • Data writt en to the same file each time Data written to the same file each increased update rate. test with one radar for increased update rate. f0 = 45 Mode = 2 1: Passenger car test subsequent data collection is stopped time subsequent data collection is • Code rewri Code rewritten tt en to crea to create te new output new sensor and stand- f = 45 with one radar sensor BW = 20 0 0 and started. stopped and started. output files upon each logging files upon each logging start. BW = 200 and standard ard graphical user Data are only written to file upon Ns = 200 • Data are only writt en to file upon log- start. N = 200 graphical user logging stop. Susceptible to data • Output files are now named with interface (GUI) ging stop. Susceptible to data loss if Ntar = 5 Output files are now named with N = 5 interface (GUI) tar loss if equipment malfunctions. “start logging time” in human- reada- equipment malfunctions. Cannot dis- “start logging time” in human- Rmax = 100 R = 100 max Cannot distinguish driving ble form readable at. format. tinguish driving segments. MTI = 1 MTI = 1 segments. Data are now continuously Mth = 1• Epoch time format impractical and • Data are now continuously writt en to Mth = 1  Epoch time format impractical written to file during logging, as not human-readable. and not human-readable. file during logging, as opposed to opposed to batch writing upon batch writing upon logging termina- logging termination. tion. 2: Test with two Mode = 2, 3 • The preceding truck was recorded • Field study involves three sensors, dump trucks. Fol- f0 = 45 well, except in curves and intersec- not one. lower with BW = 200 tions. Appl. Syst. Innov. 2023, 6, 55 23 of 29 Table A2. Cont. Supplier Modifications and Pre-Test Steps Parameters User Experience Recommendations Appl. Syst. Innov. 2023, 6, x FOR PEER REVIEW 23 of 29 The preceding truck was recorded well, except in curves and Mode = 2, 3 2: Test with two dump intersections.  Field study involves three sensors, f = 45 trucks. Follower with Mode 3 is preferred over mode 2; not one. dashcam and one Ns = 200 BW = 200 • Mode 3 is preferred over mode 2;• Finalized parameters were recom- dashcam and one makes it easy to filter away  Finalized parameters were N = 200 radar sensor Ntar = 5 makes it easy to filter away stationary mended to avoid interference, yet radar sensor stationary objects and oncoming recommended to avoid N = 5 tar Rmax = 100 objects and oncoming vehicles based maximize sensor range. vehicles based on relative velocity. interference, yet maximize sensor R = 100 max MTI = 1 on relative velocity. Filtering distance values for range. MTI = 1 relative velocity exceeding Mth = 0 Mth = 0 • Filtering distance values for relative 20 km/h removes much noise. velocity exceeding ± 20 km/h removes much noise. Table A3. Radar parameter configurations. Table A3. Radar parameter configurations. Truck ID Common Parameters Specific Parameters Truck ID Common Parameters Specific Parameters Mode = 3 f = 5 MHz Mode = 3 BW = 200 MHz f0 = 5 MHz N = 200 samples BW = 200 MHz N = 5 targets tar Ns = 200 samples Ntar = 5 targets R = 100 m max f = 25 MHz MTI = 1 (active) 3 Rmax = 100 m N = 195 samples f0 = 25 MHz Mth = 0 (inactive) MTI = 1 (active) Ns = 195 samples Mth = 0 (inactive) Appendix A.2. Video Footage, Video Synchronization and Manual Video Coding Appendix A.2. Video Footage, Video Synchronization and Manual Video Coding Low-resolution video (LRV) file segments were converted to the MP4 format and merged using free Bandicut software [47]. Merged videos were imported to BORIS version Low-resolution video (LRV) file segments were converted to the MP4 format and 7.12.2. LRV files were used, since original MP4 files were too large for BORIS to handle. LRV merged using free Bandicut software [47]. Merged videos were imported to BORIS version files were 864 by 480 pixels, while original files were 1920 by 1080 pixels, both with 60 fps 7.12.2. LRV files were used, since original MP4 files were too large for BORIS to handle. frame rates. Conversion reduced the file size by an order of magnitude, while retaining LRV files were 864 by 480 pixels, while original files were 1920 by 1080 pixels, both with sufficient video quality for coding. Table A4 shows two code definitions, while examples of 60 fps frame rates. Conversion reduced the file size by an order of magnitude, while re- state events are shown in Table A5. taining sufficient video quality for coding. Table A4 shows two code definitions, while examples of state events are shown in Table A5. Table A4. Examples of video code definitions. Table A4. Examples of video code definitions. Video Code Definition Driving starts when the truck is fully inside the correct Video Code Definition lane on the roadway, with the steering wheel turned Driving starts when the truck is fully inside the correct lane on the roadway, with the steering wheel straight. It stops just before the driver starts turning the turned straight. It stops just before the driver starts turning the wheel, with the intention of entering Driving (S) wheel, with the intention of entering driveways, parking Driving (S) driveways, parking areas or stop pockets. Except during Break and periods of camera malfunctions, areas or stop pockets. Except during Break and periods of every other video code is coded only when Driving camera is also malfunctions, active. every other video code is coded only when Driving is also active. All time that is not Driving, is defined as Break. This includes maneuvering in and out of driveways, Break (S) All time that is not Driving, is defined as Break. This parking areas and stop pockets. Break (S) includes maneuvering in and out of driveways, parking areas and stop pockets. Table A5. Examples of coded state events for radar data filtering. Event Illustrations Tunnel Appl. Appl. Syst Syst . Innov. . Innov. 2023 2023 , 6 , 6 , x FO , x FO R P R P EER R EER R EVIE EVIE W W 23 of 23 of 29 29 dashcam and one Ns = 200 dashcam and one Ns = 200 • • Mode Mode 3 is p 3 is preferr referred ov ed ov erer mode 2; mode 2; • • Finalize Finalized parameters w d parameters were recom- ere recom- rad rad ar sensor ar sensor NN tartar = = 5 5 ma ma kes it kes it ea ea sy sy to to fifi ltler aw ter aw ay sta ay sta tit o ina ona ry ry mended to mended to avoid avoid inter interfefreence, y rence, yetet Rmax = 100 Rmax = 100 objects and objects and o o nn coming v coming v eh eh icles b icles b ase ase dd ma ma ximiz ximiz e sensor ra e sensor ra nge. nge. MTI MTI = 1 = 1 on relative velocity. on relative velocity. Mth = 0 Mth = 0 • • FiFi ltleri teri nn g di g di stst an an ce v ce v alu alu es for rela es for rela titi ve v e ve vlocit elocit yy exce exce ed ed ing ing ± ± 20 2 km/ 0 km/ hh remov remov es es much noise. much noise. Ta Ta bb le l A e A 3.3 Radar parameter con . Radar parameter con figurations figurations . . Truck ID Truck ID Common Common PP ar aa rmeters ameters Spec Spec ifi ifi c P c P arame arame te te rs r s Mode = 3 Mode = 3 f0f = 0 = 5 M 5 M H H z z 2 BW = 20 BW = 20 0 MH 0 MH z z NN s = 200 s = 200 sa sa m m pp lels es N N tar tar = 5 = 5 ta trget arget s s RR max max = 10 = 10 0 0 m m f0 = 25 MHz f0 = 25 MHz 3 3 MTI MTI = 1 ( = 1 ( act act iv iv e) e) NN s = 195 s = 195 sa sa m m pp lels es Mt Mt h h = 0 (in = 0 (in act act iv ie v) e ) Ap Ap pen pen did xi x A A .2..2 V . V ideo Foota ideo Foota ge, ge, V V ideo Sy ideo Sy nc nc hro hro nn izati izati on on an an d Manual d Manual V V ideo ideo C C odi odi ng ng Low-reso Low-reso lut lut ion video ion video (LR (LR VV ) ) fifi le le segme segme nn ts w ts w ere con ere con verted to the MP verted to the MP 4 form 4 form atat and and merged usin merged usin g free B g free B and and icut icut softw softw are are [47]. Merged v [47]. Merged v ide ide os w os w ere impo ere impo rted to BORIS v rted to BORIS v ersion ersion 7.7. 12 12 .2.2 . L . L RR VV fi fi les les w w ere ere use use dd , s , s ince o ince o rig rig in in alal MP MP 4 4fi fi lele s s ww ere t ere t oo oo lar lar ge for ge for BO BO RIS RIS t t o hand o hand le. le. LRV files were 864 by 480 pixels, while original files were 1920 by 1080 pixels, both with LRV files were 864 by 480 pixels, while original files were 1920 by 1080 pixels, both with 60 fps fr 60 fps fr ame rates. ame rates. Conv Conv er er sion sion red red uu ced ced the the fifi lele s s ize b ize b yy an o an o rder rder of of ma ma gnit gnit ud ud e,e, while while re re - - tatinin ainin g su g suffiffi cicent vid ient vid eo eo quality quality for co for co ding. ding. Ta Tble able AA 4 4shows two shows two c c ode de ode defifi nn itiitoin on s, wh s, wh ile ile example example s of state ev s of state ev ents ar ents ar e shown e shown in T in T able able AA 5. 5. Ta Ta bb le l A e A 4.4 Ex . Ex am am ples of ples of vi vi deo deo code code de de finition finition s. s. Vi Vi dd eo eo C C od od e eD D efi efi nition nition Driv Driv ing ing starts when th starts when th e e truck is fully truck is fully inside th inside th e correc e correc t lane on t lane on the the r r oadw oadw ay, with ay, with the the steerin steerin g g wheel wheel turne turne dd s s trta ria g iht. I ght. I t s t s top top s ju s ju st b st b efore efore the the driv driv er er stst ar ar tsts t t uu rning rning th th e w e w hh eel, eel, wi wi th th the the int int ent ent ion o ion o f e f e nn terin terin gg Driv Driv ing ing (S (S ) ) dri dri vv ewa ewa yy s,s, pa pa rki rki nn gg a r aea rea s s or st or st op pock op pock etet s.s Except . Except du du riri ng ng Br Br eak eak and periods of camera m and periods of camera m alfun alfun ction ction s, s, every o every o ther v ther v ideo cod ideo cod e is c e is c oded only wh oded only wh en en Driv Driv ing ing is i a s a lslo s a o a cti cti ve. ve. Appl. Syst. Innov. 2023, 6, 55 24 of 29 Al Al l ti l ti me t me t hh at a is t is not not Dri Dri ving ving , is defined , is defined as as Break Break . Th . Th is is in in clude clude s m s m ane ane uvering uvering in in an an d ou d ou t o t o f dr f dr ive ive w w ays ays , , Break Break (S (S ) ) pa pa rk rk in in gg a r aea rea s a s a nn d st d st op pock op pock etet s. s. Table A5. Examples of coded state events for radar data filtering. Ta Ta bb le l A e A 5.5 Exa . Exa m m ples of ples of cod cod ed e state d state ev ev ents fo ents fo r radar data r radar data fi lt fiering. ltering. Event Illustrations EvEvenent Il t Illuluststraratitions ons Tunnel Tunnel Tunnel Appl. Appl. Syst Syst . Innov. . Innov. 2023 2023 , 6 , 6 , x FO , x FO R P R P EER R EER R EE VIE VIE W W 24 of 24 of 29 29 Appl. Appl. Syst Syst . Innov. . Innov. 2023 2023 , 6 , 6 , x FO , x FO R P R P EER R EER R EE VIE VIE W W 24 of 24 of 29 29 Appl. Syst. Innov. 2023, 6, x FOR PEER REVIEW 24 of 29 Appl. Syst. Innov. 2023, 6, x FOR PEER REVIEW 24 of 29 Roundabout Roundabout Roundabout Roundabout Roundabout Roundabout Roundabout (Left): (Left (Left Straight; ): Straigh ): Straigh t; t; (Left (Left ): Straigh ): Straigh t; t; (Right) (Left Right-turn ): Straight; (Left): Straight; (R (R ight ight ) R ) R ight ight -t-u tu rn rn (R (R ight ight ) R ) R ight ight -t-u tu rn rn (Right) Right-turn (Right) Right-turn LWRR (likely LWRR (likely with with in in LWRR (likely LWRR (likely with with in in LWRR (likely within LWRR (likely within LWRR radar r radar r (likelyawithin nge) ange) radar range) radar r radar r ange) ange) radar r radar r ange) ange) LBR LBR RR (l (l ik iel kel y b y b eyon eyon dd LBR LBR RR (l (l ik iel kel y b y b eyon eyon dd LBRR (likely beyond LBRR (likely beyond radar r radar r ange) ange) LBRR (likely beyond radar range) radar r radar r ange) ange) radar range) radar range) Ap Ap pen pen dd ix ix A A .3 .3 . .R R adar Da adar Da ta C ta C uu rati rati on on Ap Ap pen pen dd ix ix A A .3 .3 . .R R adar Da adar Da ta C ta C uu rati rati on on Appendix A.3. Radar Data Curation Appendix A.3. Radar Data Curation Ra Ra dar dar d d aa ta ta p p oo st st -process -process in in g w g w aa s c s c aa rr rr ied ied out out in in Pyt Pyt hon hon 3. 3. 11 00 . The . The Pyt Pyt hh on libr on libr aa rires ies PP aa n- n- Appendix A.3. Radar Data Curation Ra Ra dar dar d d aa ta ta p p oo st st -process -process in in g w g w aa s c s c aa rr rr ied ied out out in in Pyt Pyt hon hon 3. 3. 11 00 . The . The Pyt Pyt hh on libr on libr aa rires ies PP aa n- n- Ra Ra dar dar d d aa ta ta p p oo st st -process -process in in g w g w aa s c s c aa rr rr ied ied out out in in Pyt Pyt hon hon 3. 3. 11 00 . The . The Pyt Pyt hh on libr on libr aa rires ies PP aa n- n- das, das, NumPy NumPy , datetime, Mat , datetime, Mat pp lotlib and lotlib and oo pp enpyxl enpyxl w w ee re re used used . All r . All r aa dd aa r d r d aa ta w ta w ee re re extracted extracted das, das, NumPy NumPy , datetime, Mat , datetime, Mat pp lotlib and lotlib and oo pp enpyxl enpyxl w w ee re re used used . All r . All r aa dd aa r d r d aa ta w ta w ee re re extracted extracted Radar data post-processing was carried out in Python 3.10. The Python libraries Pan- das, das, NumPy NumPy , datetime, Mat , datetime, Mat pp lotlib and lotlib and oo pp enpyxl enpyxl w w ere ere used used . All r . All r aa dd aa r d r d aa ta w ta w ere ere extracted extracted int intoo a P a P aand ndaas Dat s DataaFram Fram e.e. In mult In multi-t i-t ararget get scenar scenarios ios, object , objects w s w eerere place placedd suc suc cessiv cessiv eely ly das, NumPy int intoo a P a P , datetime, aand ndaas Dat s Dat Matplotlib aaFram Fram e.e. In mult In mult and openpyxl i-t i-t ararget get wer scenar scenar e used. ios ios,All object , object radar s w s w data eerere place wer place e extracted dd suc suc cessiv cessiv eely ly into a Pandas DataFrame. In multi-target scenarios, objects were placed successively into a Pandas DataFrame. In multi-target scenarios, objects were placed successively wi wi thi thi nn the Data the Data FF ra ra m m e, by de e, by de sc se ce nn dd ing SNR ing SNR . The Da . The Da ta ta Fr Fr aa m m e c e c oo ntai ntai ne ne dd the f the f oo lllo lo wing wing da da ta ta into a wi Pandas wi thi thi nn the Data the Data DataFrame. FF ra ra m In m e, by de e, by de multi-tar sc se c ge e nn d td iscenarios, ng SNR ing SNR . The Da . The Da objectsta wer ta Fr Fr a ea m placed m e c e c oo ntai ntai successively ne ne dd the f the f oo lwithin llo lo wing wing da da ta ta within the DataFrame, by descending SNR. The DataFrame contained the following data within the DataFrame, by descending SNR. The DataFrame contained the following data columns: Tim columns: Tim ee { { date date time time }, Dist }, Dist ance ance { fl {fl oa oa t}, t}, V V ee locit locit yy { {flfl oat oat },}, S S N N RR { {flfl oa oa t} t} and and Objec Objec t number t number the Dacolumns: Tim tcolumns: Tim aFrame, by descending ee { { date date time time }, Dist SNR. }, Dist The ance ance DataFrame { fl {fl oa oa t}, t}, V V ee locit contained locit yy { {flfl oat oat }, the }, S S N following N RR { {flfl oa oa t} t} and data and Objec Objec columns: t number t number columns: Time {datetime}, Distance {float}, Velocity {float}, SNR {float} and Object number columns: Time {datetime}, Distance {float}, Velocity {float}, SNR {float} and Object number {int {int }. Since r }. Since r aa dar timest dar timest am am ps did not cor ps did not cor respond respond t t oo lo lo cal t cal t ime ime , dat , dat ee titme shift ime shift s w s w ee re calc re calc u- u- Time {{datetime int {int }. Since r }. Since r }, Distance aa dar timest dar timest {float am }, am V ps did not cor ps did not cor elocity {float}, respond r SNR espond {float t t oo lo } lo and cal t cal t Ob ime ime ject , dat , dat number ee titme shift ime shift {int}. s w s w Since ee re calc re calc u- u- {int}. Since radar timestamps did not correspond to local time, datetime shifts were calcu- lat {int lat e}. Since r e dd ba ba sed sed o a o dar timest nn t t hh e prev e prev am iou iou ps did not cor sly sly correct correct ed ed rdate espond date an an d t d time time o lo of cal t of each ieach me , dat Rad Rad a er logg a tir logg me shift ing ing in in s w stst a ea n re calc n ce, ce, and and u- radar timestamps did not correspond to local time, datetime shifts were calculated based lat lat ee dd ba ba sed sed o o nn t t hh e prev e prev iou iou sly sly correct correct ed ed date date an an dd time time of of each each Rad Rad ar logg ar logg ing ing in in stst aa nn ce, ce, and and lat lat ed ed ba ba sed sed o o nn t t hh e prev e prev iou iou sly sly correct correct ed ed date date an an dd time time of of each each Rad Rad ar logg ar logg ing ing in in stst aa nn ce, ce, and and the the da da tete an an dd ti ti me me wi wi th whi th whi ch the ch the ra ra da da r output r output fifi les les w w ee re n re n aa med med (c (c f.f. T T aa ble ble A A 22 ).). Dat Dat ee titme ime on the previously corrected date and time of each Radar logging instance, and the date and the the da da tete an an dd ti ti me me wi wi th whi th whi ch the ch the ra ra da da r output r output fifi les les w w ee re n re n aa med med (c (c f.f. T T aa ble ble A A 22 ).). Dat Dat ee titme ime the date and time with which the radar output files were named (cf. Table A2). Datetime the date and time with which the radar output files were named (cf. Table A2). Datetime shifts w shifts w ee re re added to the T added to the T ime column, ime column, correct correct ing ing all all measurement measurement s. s. A A ll r ll r aa dd aa r d r d aa ta w ta w ee re re time with which the radar output files were named (cf. Table A2). Datetime shifts were shifts w shifts w ee re re added to the T added to the T ime column, ime column, correct correct ing ing all all measurement measurement s. s. A A ll r ll r aa dd aa r d r d aa ta w ta w ee re re shifts were added to the Time column, correcting all measurements. All radar data were shifts were added to the Time column, correcting all measurements. All radar data were merged merged into into one Dat one Dat aa Fr Fr ame and ame and sav sav ee dd as a P as a P ickle ickle fifi le le pr pr io io r to cu r to cu ra ra titi on. on. added to the Time column, correcting all measurements. All radar data were merged into merged merged into into one Dat one Dat aa Fr Fr ame and ame and sav sav ee dd as a P as a P ickle ickle fifi le le pr pr io io r to cu r to cu ra ra titi on. on. merged into one DataFrame and saved as a Pickle file prior to curation. merged into one DataFrame and saved as a Pickle file prior to curation. Filt Filt er er s s 4 (t 4 (t arg arg ee t sect t sect ion) ion) an an d 5 d 5 (down (down sa sa m m pling pling ) )are are ilil lust lust ra ra ted ted in F in F igur igur ee s s A2 A2 an an d d A A 33 . . one DataFrame and saved as a Pickle file prior to curation. Filt Filt er er s s 4 (t 4 (t arg arg ee t sect t sect ion) ion) an an d 5 d 5 (down (down sa sa m m pling pling ) )are are ilil lust lust ra ra ted ted in F in F igur igur ee s s A2 A2 an an d d A A 33 . . Filters 4 (target section) and 5 (downsampling) are illustrated in Figures A2 and A3. Fig Filt Fig u er u re s re 4 (t A2 A2 arg sh sh ows a e ows a t sect7-min ion) 7-min an period d 5 period (down for truck 2. Chosen for truck 2. Chosen sampling) are illust targets targets rated are b are b in F lue, d ilgur ue, d eis sc isc A2 arde arde an d d d ones ones A3. Filters 4 (target section) and 5 (downsampling) are illustrated in Figures A2 and A3. Fig Fig uu re re A2 A2 sh sh ows a ows a 7-min 7-min period period for truck 2. Chosen for truck 2. Chosen targets targets are b are b lue, d lue, d isc isc arde arde dd ones ones Fig Fig uu re re A2 A2 sh sh ows a ows a 7-min 7-min period period for truck 2. Chosen for truck 2. Chosen targets targets are b are b lue, d lue, d isc isc arde arde dd ones ones purple, purple, and t and t hh e moving e moving av av erage erage t t uu rq rq uo uo is ie. se. A A vv er er ag ag ing ing acro acro ss ss 1 1 00 log log gg ings ings , inst , inst ee aa dd of of few few ee r,r , purple, purple, and t and t hh e moving e moving av av erage erage t t uu rq rq uo uo is ie. se. A A vv er er ag ag ing ing acro acro ss ss 1 1 00 log log gg ings ings , inst , inst ee aa dd of of few few ee r,r , purple, purple, and t and t hh e moving e moving av av erage erage t t uu rq rq uo uo is ie. se. A A vv er er ag ag ing ing acro acro ss ss 1 1 00 log log gg ings ings , inst , inst ea ea dd of of few few er e,r , reduced reduced noise in noise in turbulen turbulen t situations. I t situations. I nn spec spec tion tion oo f rad f rad aa r dat r dat aa , e , e .g., .g., when when trav trav ersing ersing reduced reduced noise in noise in turbulen turbulen t situations. I t situations. I nn spec spec tion tion oo f rad f rad aa r dat r dat aa , e , e .g., .g., when when trav trav ersing ersing reduced noise in turbulent situations. Inspection of radar data, e.g., when traversing reduced noise in turbulent situations. Inspection of radar data, e.g., when traversing Roundabouts Roundabouts , , show show ed that ed that approxim approxim ately 10 ately 10 logg logg ing ing s w s w ee re nee re nee ded ded after the vid after the vid ee o cod o cod ee Roundabouts Roundabouts , , show show ed that ed that approxim approxim ately 10 ately 10 logg logg ing ing s w s w ee re nee re nee ded ded after the vid after the vid ee o cod o cod ee Roundabouts, showed that approximately 10 loggings were needed after the video code Roundabouts, showed that approximately 10 loggings were needed after the video code ended, ended, for for distance to the p distance to the p recedin recedin gg truc truc k to stab k to stab ilize ilize . . ended, ended, for for distance to the p distance to the p recedin recedin gg truc truc k to stab k to stab ilize ilize . . ended, for distance to the preceding truck to stabilize. ended, for distance to the preceding truck to stabilize. The presence The presence of of lone lone blue blue d d aa ta points loc ta points loc aa ted ted aw aw ay ay fro fro m m the b the b lue lue tre tre nn d in d in Fig Fig uu re re A2, A2, The presence The presence of of lone lone blue blue d d aa ta points loc ta points loc aa ted ted aw aw ay ay fro fro m m the b the b lue lue tre tre nn d in d in Fig Fig uu re re A2, A2, The presence of lone blue data points located away from the blue trend in Figure A2, shows tha shows tha The presence t thi t thi s sal al gori of gori lone thm thm blue ma ma y choose the wr d y choose the wr ata points locaong ted ong tar aw tar ay gg et. Er et. Er fromroneous se the b roneous se lue tre lection lection nd in sFig sare those are those ure A2, shows tha shows tha t thi t thi s sal al gori gori thm thm ma ma y choose the wr y choose the wr ong ong tar tar gg et. Er et. Er roneous se roneous se lection lection s sare those are those shows tha shows tha t thi t thi s sal al gori gori thm thm ma ma y choose the wr y choose the wr ong ong tar tar gg et. Er et. Er roneous se roneous se lection lection s sare those are those data points w data points w hh ich clear ich clear ly ly suppress the suppress the moving moving av av erage d erage d istance istance v v aa lue. Figure lue. Figure A3 A3 shows shows data points w data points w hh ich clear ich clear ly ly suppress the suppress the moving moving av av erage d erage d istance istance v v aa lue. Figure lue. Figure A3 A3 shows shows data points which clearly suppress the moving average distance value. Figure A3 shows data points which clearly suppress the moving average distance value. Figure A3 shows a a 10 10 -min -min exce exce rpt rpt of downs of downs aa mpled mpled ra ra da da r dat r dat aa (b (b lue lue ), ), alongside alongside dat dat aa after after fifi ltlt er er 1 1 ( ( gg ray ray ).) . a a 10 10 -min -min exce exce rpt rpt of downs of downs aa mpled mpled ra ra da da r dat r dat aa (b (b lue lue ), ), alongside alongside dat dat aa after after fifi ltlt er er 1 1 ( ( gg ray ray ).) . a 10-min excerpt of downsampled radar data (blue), alongside data after filter 1 (gray). a 10-min excerpt of downsampled radar data (blue), alongside data after filter 1 (gray). Appl. Syst. Innov. 2023, 6, 55 25 of 29 Figure A2 shows a 7-min period for truck 2. Chosen targets are blue, discarded ones purple, and the moving average turquoise. Averaging across 10 loggings, instead of fewer, reduced noise in turbulent situations. Inspection of radar data, e.g., when traversing Roundabouts, showed that approximately 10 loggings were needed after the video code ended, for distance to the preceding truck to stabilize. The presence of lone blue data points located away from the blue trend in Figure A2, shows that this algorithm may choose the wrong target. Erroneous selections are those Appl. Syst. Innov. 2023, 6, x FOR PEER REVIEW 25 of 29 data points which clearly suppress the moving average distance value. Figure A3 shows a Appl. Syst. Innov. 2023, 6, x FOR PEER REVIEW 25 of 29 10-min excerpt of downsampled radar data (blue), alongside data after filter 1 (gray). Figure A2. Truck 2 excerpt. Target selection when radar measures multiple objects simultaneously. Figure A2. Truck 2 excerpt. Target selection when radar measures multiple objects simultaneously. Figure A2. Truck 2 excerpt. Target selection when radar measures multiple objects simultaneously. Figure A3. Truck 2 excerpt following filters 1 and 5. Figure A3. Truck 2 excerpt following filters 1 and 5. Figure A3. Truck 2 excerpt following filters 1 and 5. Appendix A.4. Results and Discussion Appendix A.4. Results and Discussion Appendix Tabl A.4. es A6 Results –A13 show and Discussion statistics from each filtering step, the analysis of relative speed Tables A6–A13 show statistics from each filtering step, the analysis of relative speed and SNR as a function of distance bins, and radar operation in different driving situations. Tables A6–A13 show statistics from each filtering step, the analysis of relative speed and SNR as a function of distance bins, and radar operation in different driving situations. and SNR as a function of distance bins, and radar operation in different driving situations. Table A6. Relative velocity statistics (km/h) for trucks 2 and 3 after each filtering step. Table A6. Relative velocity statistics (km/h) for trucks 2 and 3 after each filtering step. Truck Number Truck Number Filtering Step 2 3 Filtering Step 2 3 Min Avg Max Std Min Avg Max Std Min Avg Max Std Min Avg Max Std Raw −280.9 −5.7 281.0 40.4 −280.7 −9.4 280.7 45.9 Raw −280.9 −5.7 281.0 40.4 −280.7 −9.4 280.7 45.9 1 −280.9 −7.7 281.0 43.7 −280.7 −11.3 280.7 48.9 1 −280.9 −7.7 281.0 43.7 −280.7 −11.3 280.7 48.9 2 −29.9 −0.5 29.9 5.5 −29.9 −0.5 29.9 5.5 2 −29.9 −0.5 29.9 5.5 −29.9 −0.5 29.9 5.5 3 −29.9 −0.3 29.9 4.8 −29.9 −0.3 29.7 4.6 3 −29.9 −0.3 29.9 4.8 −29.9 −0.3 29.7 4.6 4 −29.9 −0.2 29.9 4.4 −29.9 −0.2 29.7 4.2 4 −29.9 −0.2 29.9 4.4 −29.9 −0.2 29.7 4.2 5 −29.6 −0.2 28.6 4.3 −29.9 −0.2 27.2 4.0 5 −29.6 −0.2 28.6 4.3 −29.9 −0.2 27.2 4.0 6 −29.6 −0.1 28.6 4.2 −29.7 −0.1 27.2 3.8 6 −29.6 −0.1 28.6 4.2 −29.7 −0.1 27.2 3.8 Appl. Syst. Innov. 2023, 6, 55 26 of 29 Table A6. Relative velocity statistics (km/h) for trucks 2 and 3 after each filtering step. Truck Number Filtering Step 2 3 Min Avg Max Std Min Avg Max Std Raw 280.9 5.7 281.0 40.4 280.7 9.4 280.7 45.9 1 280.9 7.7 281.0 43.7 280.7 11.3 280.7 48.9 2 29.9 0.5 29.9 5.5 29.9 0.5 29.9 5.5 3 29.9 0.3 29.9 4.8 29.9 0.3 29.7 4.6 4 29.9 0.2 29.9 4.4 29.9 0.2 29.7 4.2 5 29.6 0.2 28.6 4.3 29.9 0.2 27.2 4.0 6 29.6 0.1 28.6 4.2 29.7 0.1 27.2 3.8 Table A7. SNR statistics (dB) for trucks 2 and 3 after each filtering step. Truck Number Filtering Step 2 3 Min Avg Max Std Min Avg Max Std Raw 6.8 24.6 53.9 10.8 6.6 22.6 51.4 9.7 1 6.9 27.6 53.9 10.6 6.7 24.2 51.4 9.9 2 6.9 29.9 53.9 9.9 6.8 28.1 51.4 8.5 3 15.1 31.8 53.9 8.3 15.1 29.5 51.4 7.3 4 15.1 32.9 53.9 7.9 15.1 30.4 51.4 7.0 5 15.1 31.6 49.7 6.9 15.1 29.4 47.7 5.9 6 15.1 31.8 49.7 6.9 15.1 29.5 47.7 5.8 Table A8. Average relative velocity (km/h) and SNR (dB) for truck 2 as a function of distance. Avg. Relative Velocity Distance Bins Average SNR (dB) # Measurements % of Total (km/h) 0–10 0.4 30.0 204 1% 10–20 1.0 32.3 1191 6% 20–30 0.3 31.1 3327 18% 30–40 0.1 31.8 5074 27% 40–50 0.0 32.4 5519 30% 50–60 0.2 31.4 2151 12% 60–70 0.1 31.2 813 4% 70+ 0.7 32.2 191 1% Table A9. Average relative velocity (km/h) and SNR (dB) for truck 3 as a function of distance. Avg. Relative Velocity Distance Bins Average SNR (dB) # Measurements % of Total (km/h) 0–10 0.3 29.4 199 1% 10–20 1.2 29.1 1113 6% 20–30 0.4 28.6 4173 23% 30–40 0.0 29.0 5642 31% 40–50 0.1 30.2 4988 28% 50–60 0.0 31.0 1502 8% 60–70 0.9 31.6 270 2% 70+ 0.4 31.4 29 0% Appl. Syst. Innov. 2023, 6, 55 27 of 29 Table A10. Relative velocity statistics (km/h) for trucks 2 and 3 during LWRR and Tunnel. Relative Velocity Truck Avg Min Max Std 2 0.1 19.7 28.6 3.5 3 0.1 19.3 18.8 3.3 Table A11. Distance (meters) and SNR (dB) statistics for trucks 2 and 3 during LWRR and Tunnel. Distance SNR Truck Avg Min Max Std Avg Min Max Std 2 36.4 10.6 68.2 10.5 30.8 15.1 46.9 6.4 3 36.2 13.8 65.0 9.6 28.5 15.7 44.0 5.2 Table A12. Distance (meters) and SNR (dB) statistics for trucks 2 and 3 during LWRR and Roundabout. Distance SNR Truck Avg Min Max Std Avg Min Max Std 2 23.3 8.6 58.4 10.1 24.5 16.2 39.5 5.2 3 22.4 6.3 60.0 11.2 22.4 15.1 41.5 4.4 Table A13. Relative velocity statistics (km/h) for trucks 2 and 3 during LWRR and Roundabout. Relative Velocity Truck Avg Min Max Std 2 6.5 27.8 11.2 9.1 3 5.9 25.7 10.5 8.9 References 1. Tsugawa, S.; Jeschke, S.; Shladover, S.E. A Review of Truck Platooning Projects for Energy Savings. IEEE Trans. Intell. Veh. 2016, 1, 68–77. [CrossRef] 2. Eitrheim, M.H.R.; Log, M.M.; Tørset, T.; Levin, T.; Pitera, K. Opportunities and Barriers for Truck Platooning on Norwegian Rural Freight Routes. Transp. Res. Rec. 2022, 2676, 810–824. [CrossRef] 3. Horenberg, D. Applications within Logistics 4.0: A Research Conducted on the Visions of 3PL Service Providers. In Proceedings of the 9th IBA Bachelor Thesis Conference; University of Twente, The Faculty of Behavioural, Management and Social Sciences: Enschede, The Netherlands, 2017. 4. Bergenhem, C.; Shladover, S.; Coelingh, E.; Englund, C.; Tsugawa, S. Overview of Platooning Systems. In Proceedings of the 19th ITS World Congress, Vienna, Austria, 22–26 October 2012. 5. Konstantinopoulou, L.; Coda, A.; Schmidt, F. Specifications for Multi-Brand Truck Platooning. In Proceedings of the 8th International Conference on Weigh-In-Motion, Prague, Czech Republic, 20–23 May 2019. 6. Borealis. Available online: https://www.vegvesen.no/vegprosjekter/europaveg/e8borealis/ (accessed on 1 March 2023). 7. Simonsen, J. Ground-Breaking EU Project on Automated Heavy-Haul Freight Vehicles to Be Launched. ITS Norway: Oslo, Norway. Press Release, 1 October 2022. Available online: https://its-norway.no/ground-breaking-eu-project-on-automated- heavy-haul-freight-vehicles-to-be-launched/ (accessed on 3 January 2023). 8. Robinson, T.; Coelingh, E. Operating Platoons On Public Motorways: An Introduction To The SARTRE Platooning Programme. In Proceedings of the 17th World Congress on Intelligent Transport Systems, Busan, Republic of Korea, 25–29 October 2010. 9. Zhang, L.; Chen, F.; Ma, X.; Pan, X. Fuel Economy in Truck Platooning: A Literature Overview and Directions for Future Research. J. Adv. Transp. 2020, 2020, e2604012. [CrossRef] 10. Srisomboon, I.; Lee, S. Efficient Position Change Algorithms for Prolonging Driving Range of a Truck Platoon. Appl. Sci. 2021, 11, 10516. [CrossRef] 11. Hakobyan, G.; Yang, B. High-Performance Automotive Radar: A Review of Signal Processing Algorithms and Modulation Schemes. IEEE Signal Process. Mag. 2019, 36, 32–44. [CrossRef] 12. Scheiner, N.; Kraus, F.; Appenrodt, N.; Dickmann, J.; Sick, B. Object Detection for Automotive Radar Point Clouds—A Comparison. AI Perspect. 2021, 3, 6. [CrossRef] Appl. Syst. Innov. 2023, 6, 55 28 of 29 13. Ju, Y.; Jin, Y.; Lee, J. Design and Implementation of a 24 GHz FMCW Radar System for Automotive Applications. In Proceedings of the 2014 International Radar Conference, Lille, France, 13–17 October 2014; pp. 1–4. 14. Venon, A.; Dupuis, Y.; Vasseur, P.; Merriaux, P. Millimeter Wave FMCW RADARs for Perception, Recognition and Localization in Automotive Applications: A Survey. IEEE Trans. Intell. Veh. 2022, 7, 533–555. [CrossRef] 15. Bilik, I.; Longman, O.; Villeval, S.; Tabrikian, J. The Rise of Radar for Autonomous Vehicles: Signal Processing Solutions and Future Research Directions. IEEE Signal Process. Mag. 2019, 36, 20–31. [CrossRef] 16. Shnidman, D.A. Radar Detection in Clutter. IEEE Trans. Aerosp. Electron. Syst. 2005, 41, 1056–1067. [CrossRef] 17. Wang, L.; Giebenhain, S.; Anklam, C.; Goldluecke, B. Radar Ghost Target Detection via Multimodal Transformers. IEEE Robot. Autom. Lett. 2021, 6, 7758–7765. [CrossRef] 18. Ortiz, F.M.; Sammarco, M.; Costa, L.H.M.K.; Detyniecki, M. Applications and Services Using Vehicular Exteroceptive Sensors: A Survey. IEEE Trans. Intell. Veh. 2023, 8, 949–969. [CrossRef] 19. Gu, Y.; Hsu, L.-T.; Kamijo, S. Passive Sensor Integration for Vehicle Self-Localization in Urban Traffic Environment. Sensors 2015, 15, 30199–30220. [CrossRef] [PubMed] 20. Wang, F.; Zhuang, W.; Yin, G.; Liu, S.; Liu, Y.; Dong, H. Robust Inter-Vehicle Distance Measurement Using Cooperative Vehicle Localization. Sensors 2021, 21, 2048. [CrossRef] [PubMed] 21. Kim, T.-W.; Jang, W.-S.; Jang, J.; Kim, J.-C. Camera and Radar-Based Perception System for Truck Platooning. In Proceedings of the 2020 20th International Conference on Control, Automation and Systems (ICCAS), Busan, Republic of Korea, 13–16 October 2020; pp. 950–955. 22. URAD Radar for Raspberry Pi. Available online: https://urad.es/en/product/urad-radar-raspberry-pi/ (accessed on 26 January 2023). 23. Eitrheim, M.H.R.; Log, M.M.; Tørset, T.; Levin, T.; Nordfjærn, T. Driver Workload in Truck Platooning: Insights from an on-Road Pilot Study on Rural Roads. In Proceedings of the 32nd European Safety and Reliability Conference (ESREL 2022), Dublin, Ireland, 28 August–1 September 2022. 24. Knoop, V.L.; Wang, M.; Wilmink, I.; Hoedemaeker, D.M.; Maaskant, M.; Van der Meer, E.-J. Platoon of SAE Level-2 Automated Vehicles on Public Roads: Setup, Traffic Interactions, and Stability. Transp. Res. Rec. 2019, 2673, 311–322. [CrossRef] 25. Mills, D.L. Internet Time Synchronization: The Network Time Protocol. IEEE Trans. Commun. 1991, 39, 1482–1493. [CrossRef] 26. Emerald Sequoia LLC Emerald Time. Available online: https://emeraldsequoia.com/et/ (accessed on 27 January 2023). 27. Racelogic 04—VBOX Sport Logging. Available online: https://en.racelogic.support/VBOX_Motorsport/Product_Info/ Performance_Meters/VBOX__Sport/VBOX_Sport_User_Guide/04_-_VBOX_Sport_Logging (accessed on 8 March 2023). 28. Rohling, H.; Meinecke, M.-M. Waveform Design Principles for Automotive Radar Systems. In Proceedings of the 2001 CIE International Conference on Radar Proceedings (Cat No.01TH8559), Beijing, China, 15–18 October 2001; pp. 1–4. 29. Al-Hasan, T.M.; Shibeika, A.S.; Attique, U.; Bensaali, F.; Himeur, Y. Smart Speed Camera Based on Automatic Number Plate Recognition for Residential Compounds and Institutions Inside Qatar. In Proceedings of the 2022 5th International Conference on Signal Processing and Information Security (ICSPIS), Dubai, United Arab Emirates, 7–8 December 2022; pp. 42–45. 30. Murad, M.; Bilik, I.; Friesen, M.; Nickolaou, J.; Salinger, J.; Geary, K.; Colburn, J.S. Requirements for next Generation Automotive Radars. In Proceedings of the 2013 IEEE Radar Conference (RadarCon13), Ottawa, ON, Canada, 29 April–3 May 2013; pp. 1–6. 31. Ouaknine, A.; Newson, A.; Rebut, J.; Tupin, F.; Pérez, P. CARRADA Dataset: Camera and Automotive Radar with Range- Angle- Doppler Annotations. In Proceedings of the 2020 25th International Conference on Pattern Recognition (ICPR), Milan, Italy, 10–15 January 2021; pp. 5068–5075. 32. Radhityo Sardjono, D.; Suratman, F.Y. Istiqomah Human Motion Change Detection Based On FMCW Radar. In Proceedings of the 2022 IEEE Asia Pacific Conference on Wireless and Mobile (APWiMob), Bandung, Indonesia, 9–10 December 2022; pp. 1–6. 33. Erdyarahman, R.; Suratman, F.Y.; Pramudita, A.A. Contactless Human Respiratory Frequency Monitoring System Based on FMCW Radar. In Proceedings of the 2022 IEEE Asia Pacific Conference on Wireless and Mobile (APWiMob), Bandung, Indonesia, 9–10 December 2022; pp. 1–7. 34. Oberhammer, J.; Somjit, N.; Shah, U.; Baghchehsaraei, Z. 16—RF MEMS for Automotive Radar. In Handbook of Mems for Wireless and Mobile Applications; Uttamchandani, D., Ed.; Woodhead Publishing Series in Electronic and Optical Materials; Woodhead Publishing: Sawston, UK, 2013; pp. 518–549; ISBN 978-0-85709-271-7. 35. Friard, O.; Gamba, M. BORIS: A Free, Versatile Open-Source Event-Logging Software for Video/Audio Coding and Live Observations. Methods Ecol. Evol. 2016, 7, 1325–1330. [CrossRef] 36. Norwegian Public Roads Administration (NPRA) Handbook N302 Road Markings. Available online: https://viewers.vegnorm. vegvesen.no/product/859926/nb#id-8d46a74b-9c46-4790-b012-974ebf080aed (accessed on 20 February 2023). 37. Suzuki, H.; Nakatsuji, T. Dynamic Estimation of Headway Distance in Vehicle Platoon System under Unexpected Car-Following Situations. Transp. Res. Procedia 2015, 6, 172–188. [CrossRef] 38. Kim, T.; Park, T.-H. Extended Kalman Filter (EKF) Design for Vehicle Position Tracking Using Reliability Function of Radar and Lidar. Sensors 2020, 20, 4126. [CrossRef] [PubMed] 39. Jing, L.; Yanping, Z.; Xingang, Z. On the Maximum Unambiguous Range of LFMCW Radar. In Proceedings of the 2019 International Conference on Information Technology and Computer Application (ITCA), Guangzhou, China, 20–22 December 2019; pp. 92–94. Appl. Syst. Innov. 2023, 6, 55 29 of 29 40. Berthold, P.; Michaelis, M.; Luettel, T.; Meissner, D.; Wuensche, H.-J. Radar Reflection Characteristics of Vehicles for Contour and Feature Estimation. In Proceedings of the 2017 Sensor Data Fusion: Trends, Solutions, Applications (SDF), Bonn, Germany, 10–12 October 2017; pp. 1–6. 41. Macaveiu, A.; Câmpeanu, A. Automotive Radar Target Tracking by Kalman Filtering. In Proceedings of the 2013 11th International TEL Conference on Telecommunications in Modern Satellite, Cable and Broadcasting Services ( SIKS), Nis, Serbia, 16–19 October 2013; Volume 2, pp. 553–556. 42. Folster, F.; Rohling, H. Data Association and Tracking for Automotive Radar Networks. IEEE Trans. Intell. Transp. Syst. 2005, 6, 370–377. [CrossRef] 43. Scheiner, N.; Appenrodt, N.; Dickmann, J.; Sick, B. A Multi-Stage Clustering Framework for Automotive Radar Data. In Proceedings of the 2019 IEEE Intelligent Transportation Systems Conference (ITSC), Auckland, New Zealand, 27–30 October 2019; pp. 2060–2067. 44. Schlichenmaier, J.; Roos, F.; Kunert, M.; Waldschmidt, C. Adaptive Clustering for Contour Estimation of Vehicles for High- Resolution Radar. In Proceedings of the 2016 IEEE MTT-S International Conference on Microwaves for Intelligent Mobility (ICMIM), San Diego, CA, USA, 19–20 May 2016; pp. 1–4. 45. Domhof, J.; Kooij, J.F.P.; Gavrila, D.M. A Joint Extrinsic Calibration Tool for Radar, Camera and Lidar. IEEE Trans. Intell. Veh. 2021, 6, 571–582. [CrossRef] 46. Ji, Z.; Prokhorov, D. Radar-Vision Fusion for Object Classification. In Proceedings of the 2008 11th International Conference on Information Fusion, Cologne, Germany, 30 June–3 July 2008; pp. 1–7. 47. Bandicam Company Bandicut Video Cutter, Joiner and Splitter Software. Available online: https://www.bandicam.com/ bandicut-video-cutter/ (accessed on 27 January 2023). Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Applied System Innovation Multidisciplinary Digital Publishing Institute

Using Low-Cost Radar Sensors and Action Cameras to Measure Inter-Vehicle Distances in Real-World Truck Platooning

Download PDF
29 pages

Loading next page...
 
/lp/multidisciplinary-digital-publishing-institute/using-low-cost-radar-sensors-and-action-cameras-to-measure-inter-Ma2UOZctF0
Publisher
Multidisciplinary Digital Publishing Institute
Copyright
© 1996-2023 MDPI (Basel, Switzerland) unless otherwise stated Disclaimer Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. Terms and Conditions Privacy Policy
ISSN
2571-5577
DOI
10.3390/asi6030055
Publisher site
See Article on Publisher Site

Abstract

Article Using Low-Cost Radar Sensors and Action Cameras to Measure Inter-Vehicle Distances in Real-World Truck Platooning 1 , 2 1 , 3 4 1 Markus Metallinos Log * , Thomas Thoresen , Maren H. R. Eitrheim , Tomas Levin and Trude Tørset Department of Civil and Environmental Engineering, Norwegian University of Science and Technology (NTNU), NO-7491 Trondheim, Norway Research Division Strategic Analyses and Joint Systems, Norwegian Defence Research Establishment (FFI), NO-2027 Kjeller, Norway Department of Humans and Automation, Institute for Energy Technology (IFE), NO-1751 Halden, Norway Norwegian Public Roads Administration (NPRA), NO-7031 Trondheim, Norway * Correspondence: markus.log@ntnu.no Abstract: Many modern vehicles collect inter-vehicle distance data from radar sensors as input to driver assistance systems. However, vehicle manufacturers often use proprietary algorithms to conceal the collected data, making them inaccessible to external individuals, such as researchers. Aftermarket sensors may circumvent this issue. This study investigated the use of low-cost radar sensors to determine inter-vehicle distances during real-world semi-automated truck platooning on two-way, two-lane rural roads. Radar data from the two follower trucks in a three-truck platoon were collected, synchronized and filtered. The sensors measured distance, relative velocity and signal-to-noise ratio. Dashboard camera footage was collected, coded and synchronized to the radar data, providing context about the driving situation, such as oncoming trucks, roundabouts and tunnels. The sensors had different configuration parameters, suggested by the supplier, to avoid signal interference. With parameters as chosen, sensor ranges, inferred from maximum distance measurements, were approximately 74 and 71 m. These values were almost on par with theoretical calculations. The sensors captured the preceding truck for 83–85% of the time where they had the preceding truck within range, and 95–96% of the time in tunnels. While roundabouts are problematic, the sensors are feasible for collecting inter-vehicle distance data during truck platooning. Citation: Log, M.M.; Thoresen, T.; Eitrheim, M.H.R.; Levin, T.; Tørset, T. Keywords: inter-vehicle distance measurements; radar sensor; action camera; field study; automated Using Low-Cost Radar Sensors and truck platooning; rural road; manual video coding; field-of-view; tunnels; roundabouts Action Cameras to Measure Inter-Vehicle Distances in Real-World Truck Platooning. Appl. Syst. Innov. 2023, 6, 55. https://doi.org/10.3390/ asi6030055 1. Introduction Truck platooning refers to the innovative concept of wirelessly linking trucks into Academic Editor: Igor Korobiichuk convoys using adaptive cruise control (ACC). Platoons consist of one lead truck and one or Received: 5 April 2023 more following trucks. Wireless communication may enable shorter inter-vehicle distances Revised: 27 April 2023 than those currently considered safe for manually driven trucks, which are constrained by Accepted: 4 May 2023 the reaction times of human drivers [1,2]. Shorter inter-vehicle distances lead to a reduction Published: 6 May 2023 in aerodynamic drag, which may enable fuel savings and reduced emissions. Moreover, tight, automated vehicle control may unlock improvements in safety and efficiency of the road traffic system [2]. Platooning may also benefit society at large, in terms of cheaper, safer and more streamlined road freight operations. However, platooning is yet to be Copyright: © 2023 by the authors. commercially deployed and there are many unanswered questions. Licensee MDPI, Basel, Switzerland. Field studies are often used to explore the technology and are organized either by This article is an open access article transport companies [3] or truck manufacturers directly [4], or through larger, publicly distributed under the terms and funded undertakings, such as the KONVOI [1] and ENSEMBLE projects [5]. The stud- conditions of the Creative Commons ies have typically been conducted on highways. Few studies, if any, have investigated Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ platooning on challenging two-way, two-lane roads with oncoming traffic and narrow 4.0/). tunnels. In Norway, for example, large parts of the road network are subject to such issues, Appl. Syst. Innov. 2023, 6, 55. https://doi.org/10.3390/asi6030055 https://www.mdpi.com/journal/asi Appl. Syst. Innov. 2023, 6, 55 2 of 29 on which more research is needed to establish the feasibility of truck platooning. Public roads authorities govern the design and operations of the road network and may thus be important facilitators of truck platooning. As Norway is a small automotive market with conditions for automated vehicles, the Norwegian Public Roads Administration (NPRA) have taken a proactive role in trialing advanced transportation technologies, exemplified by the Borealis testbed [6] and the ongoing MODI project [7], which aims to demonstrate automated trucking between the Netherlands and Norway within 2026. In general, the shorter inter-vehicle distances are between platooning trucks, the greater are the resulting benefits in terms of fuel savings and potential road capacity improvements. Thus, the greater the stability of inter-vehicle distances over time, the more beneficial platooning will be [8]. However, during real-world driving, combinations of external traffic, road alignment and truck weight differences will influence the inter-vehicle distances, and thus also the extent of benefits unlocked [9,10]. Moreover, the extent to which truck platoons impact surrounding traffic depends partly on their total length, which is influenced by inter-vehicle distances. By implication, inter-vehicle distance data will be important for public roads authorities when regulating truck platooning, such as when deciding on which road sections platoons should be allowed, and the maximum number of trucks which can platoon together. However, such data may not be easily accessible, and even if they were, it is unclear how they should be contextualized and analyzed. In the field studies organized through the large, aforementioned platooning projects, truck manufacturers and transport companies have typically facilitated and allowed for the collection of inter-vehicle distance measurements using integrated vehicle sensors. While integrated distance sensors could also be cameras and lidars [11], radar sensors are most often used, as they are affordable, computationally simple and robust under adverse light and weather conditions [11–15]. This is despite issues with clutter [16] and ghosting [17], referring to unwanted signals that distort and interfere with the desired detection, and the detection of non-existent targets which are difficult to distinguish from real ones. If allowed, using data from integrated vehicle sensors is very convenient, as they are already collected as real-time inputs to platooning control systems, removing the need for using aftermarket distance sensors. However, manufacturers often use proprietary algorithms for data processing, defining message codes [18] to encrypt collected data. This makes external individuals unable to access them, unless authorized to do so. Smaller truck platooning field studies may not have the benefit of truck manufacturers participating as partners. Some may also seek to verify data from manufacturers, for which independent methods for collecting such data would be useful. High-precision global navigation satellite system (GNSS) receivers located in succes- sive vehicles could theoretically be used for this purpose [19,20]. Some areas, however, have road tunnels and topographical features where GNSS-based data collection methods would be subject to signal blockages [12,20]. External aftermarket radar sensors may circumvent the problem, provided they are adequately practical and accurate. Truck platooning re- search is often publicly funded, so solutions should preferably be low-cost. Since the output of aftermarket radar sensors would not be used for operative vehicle control, they do not need to be as capable, nor provide the same level of reliability as automotive-grade distance sensors, both of which drive cost and complexity. They may also be more flexible, in terms of allowing for custom placement and user adaptations. Many researchers have focused on perception, functional safety and operative control for truck platoons [21]. These studies often include cameras and radars. However, for studying the effects of truck platooning from the standpoint of roads authorities, these methods are more computationally complex than they need to be, and simple methods for estimating inter-vehicle distances from truck platooning would be useful. The current study investigated the feasibility of using Anteral universal radar (uRAD) sensors for Raspberry Pi [22] to measure inter-vehicle distances in a truck platoon on rural roads. This application represents a novel use case for this type of sensor. We propose a multi-faceted approach for collecting inter-vehicle distance data from truck Appl. Syst. Innov. 2023, 6, x FOR PEER REVIEW 3 of 29 on rural roads. This application represents a novel use case for this type of sensor. We propose a multi-faceted approach for collecting inter-vehicle distance data from truck pla- tooning field trials. It aims to provide technical details and best-practices on data collec- tion, synchronization, filtering and analysis. Dashboard cameras in each truck filmed the driving scene. This footage was used to log the timestamps of specific, recurring events, allowing for exploring sensor operation in different driving conditions. Appl. Syst. Innov. 2023, 6, 55 3 of 29 The paper addresses the research question: How can low-cost radar sensors and ac- tion cameras be used to investigate inter-vehicle distances in real-world truck platooning? platooning field trials. It aims to provide technical details and best-practices on data 2. Materials and Methods collection, synchronization, filtering and analysis. Dashboard cameras in each truck filmed This section provides details on the data collection set-up, the equipment used, and the driving scene. This footage was used to log the timestamps of specific, recurring events, the procedures for synchronization, video coding and radar data processing. allowing for exploring sensor operation in different driving conditions. The paper addresses the research question: How can low-cost radar sensors and action 2.1. Data Dollection Set-Up cameras be used to investigate inter-vehicle distances in real-world truck platooning? A truck platooning field trial was undertaken on public rural roads in northern Nor- way in the fall of 2020. This is the first study of its kind, and was also reported in [23]. 2. Materials and Methods Three drivers operated three semi-trailer trucks along a 380 km two-way, two-lane road This section provides details on the data collection set-up, the equipment used, and stretch traversing a mountainous, coastal area. The trucks were numbered 1, 2 and 3, the procedures for synchronization, video coding and radar data processing. based on the main truck order configuration. One section was traversed repeatedly with different orders. A prototype ACC system was installed, enabling the trucks to operate as 2.1. Data Dollection Set-Up a platoon when detecting a preceding truck. Data from integrated cameras and radars A truck platooning field trial was undertaken on public rural roads in northern Norway were unavailable. Aftermarket equipment was used to collect data over 7 h of driving. in the fall of 2020. This is the first study of its kind, and was also reported in [23]. Three While longitudinal control was automated, the drivers operated the wheel manually, drivers operated three semi-trailer trucks along a 380 km two-way, two-lane road stretch placing the field study at the Society of Automotive Engineers (SAE) Level 1 [24]. All traversing a mountainous, coastal area. The trucks were numbered 1, 2 and 3, based on the trucks had 500 horsepower. Trucks 1 and 2 had equal weights, while truck 3 was lighter. main truck order configuration. One section was traversed repeatedly with different orders. (41 and 27.5 metric tons), reflecting the scenario of trucks encountering each other to pla- A prototype ACC system was installed, enabling the trucks to operate as a platoon when toon during real-world operations, unharmonized with respect to weight. Consequently, detecting a preceding truck. Data from integrated cameras and radars were unavailable. inter-vehicle distances fluctuated, as the platoon often struggled to remain collected on Aftermarket equipment was used to collect data over 7 h of driving. the winding road. Twenty-three tunnels and eleven roundabouts were traversed. The While longitudinal control was automated, the drivers operated the wheel manually, most prevalent speed limits were 80 and 60 km/h, during which inter-vehicle distances at placing the field study at the Society of Automotive Engineers (SAE) Level 1 [24]. All trucks 3-second time gaps were 50–70 m, i.e., comparable to manual driving. Figure 1 illustrates had 500 horsepower. Trucks 1 and 2 had equal weights, while truck 3 was lighter. (41 and the set-up. 27.5 metric tons), reflecting the scenario of trucks encountering each other to platoon during Herein, preceding truck is a term used for the truck located in front of the truck in real-world operations, unharmonized with respect to weight. Consequently, inter-vehicle question. Depending on the context, it may refer to either the lead truck or the middle distances fluctuated, as the platoon often struggled to remain collected on the winding truck. In truck order 1-2-3, truck 1 precedes truck 2, and truck 2 precedes truck 3. The road. Twenty-three tunnels and eleven roundabouts were traversed. The most prevalent terms leader or leading truck are only used for the truck located at the front of the platoon, speed limits were 80 and 60 km/h, during which inter-vehicle distances at 3-s time gaps while followers refer to both the middle and last truck together. were 50–70 m, i.e., comparable to manual driving. Figure 1 illustrates the set-up. Figure 1. Truck platooning set-up. Radar sensors were mounted on windshields of following trucks. Figure 1. Truck platooning set-up. Radar sensors were mounted on windshields of following trucks. The two rearmost trucks were identically instrumented. The Raspberry Pi, with the Herein, preceding truck is a term used for the truck located in front of the truck in uRAD sensor att ached, was fixed to the inside of the windshield using a suction mount question. Depending on the context, it may refer to either the lead truck or the middle with a flexible arm. The mount did not interfere with the field-of-view of the sensors. A truck. In truck order 1-2-3, truck 1 precedes truck 2, and truck 2 precedes truck 3. The terms leader or leading truck are only used for the truck located at the front of the platoon, while followers refer to both the middle and last truck together. The two rearmost trucks were identically instrumented. The Raspberry Pi, with the uRAD sensor attached, was fixed to the inside of the windshield using a suction mount with a flexible arm. The mount did not interfere with the field-of-view of the sensors. A portable monitor was used to administer radar logging. Sensors were placed at slightly different heights in each truck due to interior constraints, cf. Appendix A.1. Two GoPro video cameras were also mounted: A windshield-mounted dashboard camera filmed the driving scene, and another camera filmed the interior. Footage from Appl. Syst. Innov. 2023, 6, x FOR PEER REVIEW 4 of 29 portable monitor was used to administer radar logging. Sensors were placed at slightly different heights in each truck due to interior constraints, cf. Appendix A.1. Appl. Syst. Innov. 2023, 6, 55 4 of 29 Two GoPro video cameras were also mounted: A windshield-mounted dashboard camera filmed the driving scene, and another camera filmed the interior. Footage from the latt er was only used for synchronizing radar data to the dashboard footage. The study the was a latter pprov was ed by the Norw only used for synchr egian Centre for onizing radar Research Da data to the ta (4 dashboar 57013). P d footage. articipan The ts ag study reed to being recorded, and all videos and audio were handled and stored confidentially. was approved by the Norwegian Centre for Research Data (457013). Participants agreed to The equipment in each truck was started and stopped in succession when trucks were being recorded, and all videos and audio were handled and stored confidentially. parked, but The equipment was left on during short brea in each truck was started ks. None and stopped of the Raspb in succession erry Pi when micro trp ucks rocessors were parked, had internet connecti but was left on on, so during they di short d not a breaks. dhere to l None ofothe cal ti Raspberry me. An equi Pi micr pment sta oprocessors rt-up pro- had internet connection, so they did not adhere to local time. An equipment start-up procedure cedure was devised which allowed for post-hoc synchronization of videos and radar data. was devised which allowed for post-hoc synchronization of videos and radar data. For For each truck, all cameras were started before starting radar logging. When starting each each truck, all cameras were started before starting radar logging. When starting each camera, the Emerald Sequoia Time smartphone application was presented. Using Net- camera, the Emerald Sequoia Time smartphone application was presented. Using Network work Time Protocol (NTP) servers, which synchronizes computer clocks over the internet Time Protocol (NTP) servers, which synchronizes computer clocks over the internet [25], [25], this application provides more accurate times than those typically provided by inter- this application provides more accurate times than those typically provided by internal nal clocks [26]. The application shows local time, and, when cellular reception is available, clocks [26]. The application shows local time, and, when cellular reception is available, it it calculates deviations from NTP time. The mean offset was 0.08 s, i.e., negligible. calculates deviations from NTP time. The mean offset was 0.08 s, i.e., negligible. GNSS data were collected from VBOX Sport loggers and a fleet management system GNSS data were collected from VBOX Sport loggers and a fleet management system (FMS), in an effort to compute inter-vehicle distances to validate the radar data. A script (FMS), in an effort to compute inter-vehicle distances to validate the radar data. A script was writt en to interpolate timestamps and calculate distances between GNSS locations was written to interpolate timestamps and calculate distances between GNSS locations from each truck. The loggers were supposed to activate automatically [27], but this func- from each truck. The loggers were supposed to activate automatically [27], but this func- tionality occasionally failed. GNSS files were also extracted from the FMS, and all files tionality occasionally failed. GNSS files were also extracted from the FMS, and all files were visualized in QGIS. Both systems experienced outages in tunnels. While the FMS were visualized in QGIS. Both systems experienced outages in tunnels. While the FMS had good positioning accuracy outside tunnels, its update rate was too low, and loggings had good positioning accuracy outside tunnels, its update rate was too low, and loggings were not always synchronized across the trucks. VBOX data which did get collected had were not always synchronized across the trucks. VBOX data which did get collected had frequent outages, and timestamps were often erroneous, placing trucks in incorrect order. frequent outages, and timestamps were often erroneous, placing trucks in incorrect order. This highlights the utility of radar in estimating inter-vehicle distances in such areas. This highlights the utility of radar in estimating inter-vehicle distances in such areas. 2.2. Radar Sensors 2.2. Radar Sensors Frequency modulated continuous waves (FMCWs) are radar waveforms often used Frequency modulated continuous waves (FMCWs) are radar waveforms often used to measure distances in automotive applications [13,28]. Anteral uRAD radar sensors for to measure distances in automotive applications [13,28]. Anteral uRAD radar sensors for Raspberry Pi were tested here, shown in Figure 2. These are 24 GHz FMCW radar sensors Raspberry Pi were tested here, shown in Figure 2. These are 24 GHz FMCW radar sensors which connect conveniently as extension boards to Raspberry Pi microprocessors [29]. which connect conveniently as extension boards to Raspberry Pi microprocessors [29]. Such Such microprocessors run a user-friendly operating system and can interface with pur- microprocessors run a user-friendly operating system and can interface with purpose- pose-built components. The automotive industry is increasingly using 77 GHz radar sen- built components. The automotive industry is increasingly using 77 GHz radar sensors, sors, allowing for increased range resolution and accuracy [30]. These sensors can bett er allowing for increased range resolution and accuracy [30]. These sensors can better separate separate closely spaced objects, and can be packaged in a smaller form factor. However, closely spaced objects, and can be packaged in a smaller form factor. However, 24 GHz 24 GHz sensors are less expensive, and automotive-grade 77 GHz sensors which could be sensors are less expensive, and automotive-grade 77 GHz sensors which could be operated operated from Raspberry Pi microprocessors were not available when procuring the from Raspberry Pi microprocessors were not available when procuring the equipment. equipment. Thus, testing the proposed methodology using cheaper 24 GHz sensors was Thus, testing the proposed methodology using cheaper 24 GHz sensors was considered considered reasonable. While many 77 GHz FMCW radar sensors are more range-capable reasonable. While many 77 GHz FMCW radar sensors are more range-capable and have and have wider fields-of view, some 77 GHz radars, e.g., in [31], have shorter ranges than wider fields-of view, some 77 GHz radars, e.g., in [31], have shorter ranges than the the uRAD sensors. uRAD sensors. Figure 2. Radar sensor outline [22]. Appl. Syst. Innov. 2023, 6, 55 5 of 29 The uRAD sensors have a theoretical distance range approaching 100 m, 30 fields-of- view both horizontally and vertically, and are able to detect up to five objects simultaneously. For each object, distance, radial velocity and reflected power (SNR, i.e., signal-to-noise ratio) are registered. In the context of truck platooning, both the sensor and the desired objects are in motion; so relative velocities are detected. The velocity range (0.2 to 75 m per second) is within the range of values encountered in road traffic. The use case and intended sensor placement were described to the supplier, which expected the application to be feasible, but did note that similar tests had never been carried out before. The sensor has been used for other purposes, the most relevant of which is as a stationary speed sensor [29,32,33]. However, transferability from the cited studies is limited, due to different use cases and configurations. Eight parameters were used to configure the sensors. These are detailed in Appendix A.1, alongside pre-trial testing of the mounting set-up and parameters. The most important con- siderations are detailed here. Firstly, the radar mode details the waveforms transmitted by the sensor. Triangular waves were chosen, which maximized the sensor range and the update rate for outputs (9–13 Hz) from the radar script. This mode also allowed for subsequent data filtering based on relative velocity. Maximizing range was important for capturing data even when trucks were located far apart, as the trucks were expected to drive with human-level gap sizes (2–3 s) at distances approaching the upper distance range. Adverse road geometry would presumably also lead to safer driving at larger gap sizes. Maximizing the update rate was seen as beneficial for obtaining as many measurements as possible. The update rate of the uRAD radar sensor is comparable to the 77 GHz sensors showcased in [14]. Secondly, number of targets detected (N ) and the detection distance (R ) were tar max maximized, to capture the most data, and to enable filtering of unwanted detections later. Thirdly, moving Target Indicator (MTI) was activated, for including data only from objects with motion relative to the sensor. The supplier stated that it would only eliminate objects which were absolutely static, such as detections of the windshield. The preceding truck would still be registered, even when moving at the same velocity as the sensor. Fourthly, for each truck, different values for ramp start frequency (f ) and the duration of each wave ramp (N ) were used for each sensor to avoid interference. Since each sensor had different N and f values, their theoretical maximum distance ranges also s 0 differed, at 75.0 and 73.1 m, for trucks 2 and 3, respectively, based on Equation A3 in Appendix A.1. These values are in line with 70–75 m estimates from the supplier. For comparison, automotive radar sensors typically have ranges of 30–150 m [30,34]. The sensors had a stated distance accuracy of 0.3%, corresponding to a 0.23 m deviation at 75 m, which is considered sufficient for the current use case. Table A3 in Appendix A.1 shows all parameters which were used. 2.3. Video Footage, Synchronization and Manual Video Coding Dashboard footage was recorded for exploring the radar data as a function of the driving scene. Without the videos, this would not have been possible. First, footage had to be coded, i.e., timestamps had to be established for relevant events in the footage. This is different from the more computationally complex process of semantic segmentation used in computer vision, which involves categorizing relevant objects in the scene, often using bounding boxes [31]. Video footage was synchronized and aligned to local time using BORIS, i.e., Behavioral Observation Research Interactive Software [35]. BORIS is a free, open-source video coding program. Each BORIS observation contained videos from the same truck and driving stretch. By checking time differences as displayed by the phone application at the start of each recording, time offsets were established, achieving near-perfect synchronization. The date and time of each observation was defined as the local time shown by the phone application to the longest video file, as the recording was started. This ensured that all videos from the trucks were aligned to local time during the field trial. Appl. Syst. Innov. 2023, 6, x FOR PEER REVIEW 6 of 29 Appl. Syst. Innov. 2023, 6, 55 6 of 29 application to the longest video file, as the recording was started. This ensured that all videos from the trucks were aligned to local time during the field trial. Events in BORIS were defined using the ethnogram, and were either point or state Events in BORIS were defined using the ethnogram, and were either point or state events: Point events had no duration (i.e., having only one timestamp), while state events events: Point events had no duration (i.e., having only one timestamp), while state events did (i.e., having both start and end timestamps). Video coding was carried out while play- did (i.e., having both start and end timestamps). Video coding was carried out while ing videos at 2–4 times normal speed, depending on driving scene complexity. Events playing videos at 2–4 times normal speed, depending on driving scene complexity. Events were coded by the first author, ensuring consistency. Videos codes were subsequently re- were coded by the first author, ensuring consistency. Videos codes were subsequently viewed by the third author. Onwards, italics are used to refer to the video codes, as illus- reviewed by the third author. Onwards, italics are used to refer to the video codes, as trated in Figure 3. Radar initiations (Radar logging) and oncoming traffic were defined as illustrated in Figure 3. Radar initiations (Radar logging) and oncoming traffic were defined point events. Amongst oncoming vehicles, only Trucks seemed to affect inter-vehicle dis- as point events. Amongst oncoming vehicles, only Trucks seemed to affect inter-vehicle tances during platooning. Video footage showed that when encountering large trucks on distances during platooning. Video footage showed that when encountering large trucks on narrow road segments, the lead truck often reduced its speed, causing speed reductions narrow road segments, the lead truck often reduced its speed, causing speed reductions also also for the followers and a contraction in inter-vehicle distances. Truck order codes indi- for the followers and a contraction in inter-vehicle distances. Truck order codes indicate cate which periods the instrumented trucks collected relevant inter-vehicle distance data which periods the instrumented trucks collected relevant inter-vehicle distance data within within the platoon, as opposed to the periods where they served as platoon leaders, col- the platoon, as opposed to the periods where they served as platoon leaders, collecting lecting irrelevant data preceding the platoon. irrelevant data preceding the platoon. Figure 3. Overview of BORIS video codes. Figure 3. Overview of BORIS video codes. The Driving video code in Figure 3 includes all conditions encountered, i.e., including The Driving video code in Figure 3 includes all conditions encountered, i.e., includ- tunnels and roundabouts, thus showcasing diverse, complex driving segments. See Table ing tunnels and roundabouts, thus showcasing diverse, complex driving segments. See A5 in Appendix A.3 for its definition. Tunnel and Roundabout video codes denote scenarios Table A5 in Appendix A.2 for its definition. Tunnel and Roundabout video codes denote of particular interest. In tunnels, inter-vehicle distances between the trucks cannot be de- scenarios of particular interest. In tunnels, inter-vehicle distances between the trucks can- termined using GNSS-based methods, and it is also unclear how tunnels affect operating not be determined using GNSS-based methods, and it is also unclear how tunnels affect conditions for the radar sensors. Roundabouts are demarcated areas (i.e., they are simple operating conditions for the radar sensors. Roundabouts are demarcated areas (i.e., they to code from video footage) with small horizontal radii, which can illustrate effects of road are simple to code from video footage) with small horizontal radii, which can illustrate ef curv fects ature of r on oad rad curvatur ar operat e on ion when radar operation the preceding when truck turn the preceding s. All e tr vuck ents w turns. ere co All ded events sep- wer arately e coded for eseparately ach truck. for Tun each nels w truck. ere co Tunnels ded from the were coded moment when the front of from the moment whenthe truc the front k in question entered the tunnel, to when the front of the truck left the tunnel. The same of the truck in question entered the tunnel, to when the front of the truck left the tunnel. The principle w same principle as used for rounda was used for bouts, i.e., codi roundabouts, ng the i.e., coding moment when the front of each truck the moment when the front of each entered truck and e enter xited the circ ed and exited ulating the cir are culating a. area. Events were defined for visual inspection of the distance to (visibility of) the preceding Events were defined for visual inspection of the distance to (visibility of ) the preced- truck, as the trucks were at times located far apart. The goal was to remove data from ing truck, as the trucks were at times located far apart. The goal was to remove data from periods when the preceding truck was difficult or impossible for the sensor to detect, due periods when the preceding truck was difficult or impossible for the sensor to detect, due to the driving situation. This occurred in two scenarios. Firstly, it occurred in sharp turns, to the driving situation. This occurred in two scenarios. Firstly, it occurred in sharp turns, where the preceding truck would disappear from radar field-of-view. Dashboard cameras where the preceding truck would disappear from radar field-of-view. Dashboard cameras had larger horizontal fields-of-view than radar sensors, so when no preceding vehicle had larger horizontal fields-of-view than radar sensors, so when no preceding vehicle ap- appeared on camera, the radar would also not detect it. Secondly, it occurred when the peared on camera, the radar would also not detect it. Secondly, it occurred when the trucks drove far apart. The distance range was shorter for radar than for the dashboard trucks drove far apart. The distance range was shorter for radar than for the dashboard camera, which was only constrained by line-of-sight. Both scenarios were coded as Likely Appl. Syst. Innov. 2023, 6, 55 7 of 29 beyond radar range (LBRR). Conversely, Likely within radar range (LWRR) denotes driving periods when relevant radar data likely could have been collected. The manual nature of this process introduces some limitations. Periods when the preceding truck was actually LWRR may have been coded as LBRR, and vice versa. Tran- sitions between these codes may also occur at different distances. Albeit imperfect, this categorization is preferable versus including all radar data, even when the preceding truck was located far beyond radar range. For the far-apart scenario, centerline road markings initially aided the visual estimation. On rural roads with 80–90 km/h speed limits, these markings have standardized lengths and gaps totaling 12 m, which repeat continuously, cf. pp. 22 in the Norwegian road marking design manual [36]. After having coded LWRR and LBRR using road markings for some time, the remaining dashboard camera footage was coded without conscious reference to the road markings. It was also attempted to use pixel counts of the preceding truck for this purpose, but doing so at large scale was unsuccessful. An overview of the data collection and processing steps is provided in Table 1. See Table A5 in Appendix A.2 for examples of video codes. Table 1. Overview of the data collection and processing steps. Step Context Description Start GoPro-cameras successively, while, for each camera, presenting local time on phone screen. Equipment start-up and logging 2 Start radar logging script while producing loud verbal cue. 3 Data collection Platoon driving. Stop GoPro camera recordings successively. Stop radar 4 Equipment logging stop logging. 5 Data transfer Import GoPro video files and raw radar files to computer. For each truck: Synchronize GoPro video footage in BORIS, using offset values. Synchronization is based on the Synchronize GoPro videos with each 6 difference between local time presented to each camera other upon starting the recordings, and fine-tuned using recorded audio. Define Date and time in BORIS observation equal to the local time shown to the reference camera (i.e., the longest video 7 Synchronize GoPro videos to local time file) by the phone application when the reference recording was started. Code Radar logging based on visual and verbal cues from 8 Video coding interior camera. Code remaining events from dashboard camera footage. 9 Export events list for each observation to spreadsheets. Apply datetime shift to radar timestamps based on Date and Synchronize radar data to local time 10 time for each BORIS observation to match them with Radar logging events. 11 Radar data curation Radar data were curated using six filters. 2.4. Radar Data Processing Video coding events were exported from BORIS as spreadsheets, and the Date and time from the corresponding observation was added to the timestamp of each instance of the Radar logging video code. This assigned local time to the instance when radar logging was started, and served as basis for synchronizing video codes and radar data in Python. Filters were needed to extract only the inter-vehicle distances between the platooning trucks. Filtering aimed at removing data from periods when the trucks were not driving (i), data which did not correspond to the preceding truck (ii), noise (iii), and finally, data from periods when the preceding truck was outside sensor range (iv). Appl. Syst. Innov. 2023, 6, 55 8 of 29 Appl. Syst. Innov. 2023, 6, x FOR PEER REVIEW 8 of 29 Timestamps, distance (m), relative velocity (km/h) and signal-to-noise ratio (decibel, (i), data which did not correspond to the preceding truck (ii), noise (iii), and finally, data dB) were logged, for up to five simultaneously detected targets. Positive relative velocities from periods when the preceding truck was outside sensor range (iv). corresponded to targets receding from the radar, and negative relative velocities corre- Timestamps, distance (m), relative velocity (km/h) and signal-to-noise ratio (decibel, sponded to approaching targets. For the curated radar dataset, an example of the former dB) were logged, for up to five simultaneously detected targets. Positive relative velocities would be the preceding truck accelerating away from the truck in question. Conversely, corresponded to targets receding from the radar, and negative relative velocities corre- the preceding truck decelerating would be an example of the latter. SNR denotes the ratio sponded to approaching targets. For the curated radar dataset, an example of the former of the signal power to the noise power. Larger and more reflective objects will produce would be the preceding truck accelerating away from the truck in question. Conversely, measurements with higher SNR values. The radar data were curated using successive the preceding truck decelerating would be an example of the latt er. SNR denotes the ratio filters, cf. Table 2. The following paragraph outlines the details and purpose of each filter. of the signal power to the noise power. Larger and more reflective objects will produce measurements with higher SNR values. The radar data were curated using successive fil- Table 2. Overview of radar data filters. ters, cf. Table 2. The following paragraph outlines the details and purpose of each filter. Filter Description Table 2. Overview of radar data filters. 1 Driving and following Filter Description 2 Relative velocity within 30 km/h 1 Driving and following 3 Signal-to-noise ratio < 15 dB 2 Relative velocity within ±30 km/h 4 Target selection 3 Signal-to-noise ratio < 15 dB 4 T 5 arget selection Downsampling 1 Hz 5 Downsampling 1 Hz 6 Likely within radar range (LWRR) 6 Likely within radar range (LWRR) First, the Driving video code was used to remove data collected during irrelevant First, the Driving video code was used to remove data collected during irrelevant pe- periods. It discards data from Break periods, so only data from Driving periods remain. riods. It discards data from Break periods, so only data from Driving periods remain. Sim- Simultaneously, truck order codes were used to exclude radar data collected during periods ultaneously, truck order codes were used to exclude radar data collected during periods when each truck served as platoon leader. Specifically, radar data from truck 2 stem from when each truck served as platoon leader. Specifically, radar data from truck 2 stem from the driving periods with truck orders 1–2–3, 1–3–2 and 3–2–1, while radar data from truck the driving periods with truck orders 1–2–3, 1–3–2 and 3–2–1, while radar data from truck 3 stem from periods coded as 1–2–3, 1–3–2 and 2–1–3. Data were discarded from periods 3 stem from periods coded as 1–2–3, 1–3–2 and 2–1–3. Data were discarded from periods where external vehicles (Other vehicle) preceded each respective truck. where external vehicles (Other vehicle) preceded each respective truck. Filters for relative velocity and SNR (filters 2 and 3, respectively), were used to clean Filters for relative velocity and SNR (filters 2 and 3, respectively), were used to clean the remaining data. Relative velocities were explored in histograms, shown in Figure 4a. the remaining data. Relative velocities were explored in histograms, shown in Figure 4a. Some relative velocity bins were far more frequent than others, giving the dataset large Some relative velocity bins were far more frequent than others, giving the dataset large dynamic range. The vertical axis is logarithmic, magnifying bins with few measurements. dynamic range. The vertical axis is logarithmic, magnifying bins with few measurements. (a) (b) Figure 4. (a) Relative velocity histograms. (b) Distance histograms with relative velocity filters. Appl. Syst. Innov. 2023, 6, 55 9 of 29 The histograms show two data spikes. Relative to a forward-facing sensor mounted in a moving truck, stationary surroundings have negative speed comparable to the speed of the truck. Assuming traffic moves at the speed limit, oncoming vehicles are measured with negative relative velocities at twice the speed limit. Similarly, measurements from the preceding truck have relative velocities fluctuating around 0 km/h. These two clusters of relative velocities appear as vertical spikes in Figure 4a. The cluster from the preceding truck was the largest. This is as expected in car-following situations, which necessitate continuous acceleration and deceleration [37]. The smaller clusters in Figure 4a had an order of magnitude fewer measurements, all of which had relative velocities in the 30 to 160 km/h range. They included static objects, oncoming traffic and measurement noise. Thus, for the relative velocity filter (filter 2), choosing 30 km/h as the lower threshold was natural, placing it at the local minimum between the spikes. Similarly, setting the upper threshold at +30 km/h made it so the entire top spike was included, while minimizing the inclusion of measurements from the noise floor. As shown in Figure 4b, the radar data were also subjected to different relative veloc- ity filters, starting at 40 km/h, successively constricting by 10 km/h steps until the narrowest filter of 10 km/h. The color of each filter reflects remaining data points after that filter has been applied. For instance, remaining data after relative velocity filtering at 30 km/h are shown in red. The 40 km/h filter left a spike from 10–20 m for truck 3. The 30 km/h filter removed most of the spike, and subsequent filter constriction did not cause notable differences. Thus, 30 km/h was chosen, striking a balance between retaining most measurements corresponding to the preceding trucks and minimizing unwanted ones (oncoming and stationary objects), while including situations with sudden braking and acceleration in the platoon, which are perhaps the most interesting ones from a safety and fuel savings perspective. The radar sensors were listed as having a velocity accuracy of 0.25 m per second. At 30 km/h cut-offs, this amounts to a possible deviation of 3%, which is considered acceptable. Still, as shown in Figure 4b, relative velocity filtering did not remove the leftmost spikes of measurements at distances too short to represent the preceding truck (0–5 m). Since the platoon drove with 2–3 s gaps, these do not represent the preceding truck. Calculations using Equation (A1), based on the 30 vertical field-of-view of the sensor, show that the road can be detected at 8–8.5 m forwards, meaning further away than the spikes. Thus, they are most likely clutter due to roadside detections, such as such as rock faces, tunnel walls, guardrails and signposts. Compared to the large, reflective rear walls of preceding trucks, such measurements should presumably be noisy, i.e., have small SNR values. Conversely, if these measurements did originate from the preceding truck, they should have had large accompanying SNR values. SNR values at short distances were indeed found to be small, and filtering for SNR < 15 dB was successful in removing them. As shown in Figure 5a, 15 dB filtering fell at a local minimum or saddle point between two distinct SNR data spikes. The data were also subjected to different SNR filters, illustrated in Figure 5b, but the filter was not constricted further, as doing so caused removal of data points with distance values around 40 m, likely corresponding to the preceding truck. Since the radar sensors were able to measure up to 5 detections simultaneously, the next filter (filter 4) involved selecting only one desired target in multi-target instances: The one most likely corresponding to the preceding truck. The distance value of each detection was compared with the average distance values of the previous 10 measurements (moving average). The detection with the smallest difference was chosen. However, occasional single-object detections had distances which were quite different from the general trend. In such cases, the algorithm had no choice but to select the only detection available. This produced spikes or drops in distance which affected the moving average. This problem was subsequently minimized by filter 5, which downsampled the data to 1 Hz (one measurement per second) by averaging all distance measurements within each second. This temporal resolution was considered sufficient, and also allowed for direct coupling between curated radar data and event durations from video footage. Appl. Syst. Innov. 2023, 6, 55 10 of 29 Appl. Syst. Innov. 2023, 6, x FOR PEER REVIEW 10 of 29 (a) (b) Figure 5. (a) SNR histograms after relative velocity filtering. (b) Distance histograms with different Figure 5. (a) SNR histograms after relative velocity filtering. (b) Distance histograms with different SNR threshold (after relative velocity filtering). SNR threshold (after relative velocity filtering). Since the radar sensors were able to measure up to 5 detections simultaneously, the Inspecting plots of distance versus time for downsampled radar data revealed the next filter (filter 4) involved selecting only one desired target in multi-target instances: The presence of sporadic periods entirely void of points, and also to periods when points were one most likely corresponding to the preceding truck. The distance value of each detection scattered (i.e., having varying distance values following no obvious trend). To understand was compared with the average distance values of the previous 10 measurements (moving these detections, video footage was coded for LWRR and LBRR. Filter 6 used these codes to average). The detection with the smallest difference was chosen. However, occasional sin- include only data collected in LWRR periods, and to exclude data collected during LBRR gle-object detections had distances which were quite different from the general trend. In periods. Cameras malfunctioned at times, during which LWRR and LBRR could not be such cases, the algorithm had no choice but to select the only detection available. This coded. Associated radar data were discarded, ensuring methodological consistency. produced spikes or drops in distance which affected the moving average. This problem In brief, filters 1 and 6 were based on manual video codes, while filters 2 and 3 were was subsequently minimized by filter 5, which downsampled the data to 1 Hz (one meas- based on recorded radar metrics, making them the most interesting ones in terms of radar urement per second) by averaging all distance measurements within each second. This operation. Filters 4 and 5 were computational heuristics. See Appendix A.3 for more details. temporal resolution was considered sufficient, and also allowed for direct coupling be- 3. Results and Discussion tween curated radar data and event durations from video footage. Inspecting plots of distanc This section explores effects e ve of rsus the time filtering for downsampled radar process, before discussing data redif vealed ferences the presence of sporadic periods entirely void of points, and also to periods when points were between expected and empirical maximum distance ranges. It also explores the ability sca of ttthe ered sensors (i.e., hav in imeasuring ng varying d the istanc preceding e values trfoll uck owing no obvi in different driving ous trend) situations. . To understa Finally nd , these detectio suggestions for ns, v futur ideo e foot work age are w made. as code See d for Appendix LWRR and A.4 LBRR for complete . Filter 6 used data tables. these codes to include only data collected in LWRR periods, and to exclude data collected during LBRR 3.1. Impacts of Filtering periods. Cameras malfunctioned at times, during which LWRR and LBRR could not be coded. Impacts Associated r on the ad number ar data w ofedata re disc points arded, arensurin e detailed, g methodolo before discussing gical consthe istency. effects on recor In brief, ded metrics: filters relative 1 and 6 w velocity ere based , signal-to-noise on manual v ratio ideo and codes, while distance. filters 2 and 3 were Figure 6 shows the sizes of datasets as a function of the filtering steps. The datasets of based on recorded radar metrics, making them the most interesting ones in terms of radar trucks 2 and 3 were affected similarly. Downsampling (filter 5) included a mean of six data operation. Filters 4 and 5 were computational heuristics. See Appendix A.3 for more de- points and a mode of seven. Figure 7 is a distance histogram showing the effects of all tails. filtering steps. Note how early filters mostly remove measurements at close distances. 3. Results and Discussion This section explores effects of the filtering process, before discussing differences be- tween expected and empirical maximum distance ranges. It also explores the ability of the sensors in measuring the preceding truck in different driving situations. Finally, sugges- tions for future work are made. See Appendix A.4 for complete data tables. Appl. Syst. Innov. 2023, 6, x FOR PEER REVIEW 11 of 29 Appl. Syst. Innov. 2023, 6, x FOR PEER REVIEW 11 of 29 3.1. Impacts of Filtering 3.1. Impacts of Filtering Impacts on the number of data points are detailed, before discussing the effects on Impacts on the number of data points are detailed, before discussing the effects on recorded metrics: relative velocity, signal-to-noise ratio and distance. recorded metrics: relative velocity, signal-to-noise ratio and distance. Figure 6 shows the sizes of datasets as a function of the filtering steps. The datasets Figure 6 shows the sizes of datasets as a function of the filtering steps. The datasets of trucks 2 and 3 were affected similarly. Downsampling (filter 5) included a mean of six Appl. Syst. Innov. 2023, 6, 55 of trucks 2 and 3 were affected similarly. Downsampling (filter 5) included a mean of 11 of six 29 data points and a mode of seven. Figure 7 is a distance histogram showing the effects of data points and a mode of seven. Figure 7 is a distance histogram showing the effects of all filtering steps. Note how early filters mostly remove measurements at close distances. all filtering steps. Note how early filters mostly remove measurements at close distances. Figure 6. Number of measurements after each filtering step has been applied. Figure 6. Number of measurements after each filtering step has been applied. Figure 6. Number of measurements after each filtering step has been applied. Figure 7. Histogram of measured distances as a function of filtering steps. Figure 7. Histogram of measured distances as a function of filtering steps. Figure 7. Histogram of measured distances as a function of filtering steps. 3.1.1. Relative Velocity Overall, filtering reduced relative velocity data spread. Still, filter 1 increased relative velocity and data spread for trucks 2 and 3. In breaks, trucks were parked behind one another. Thus, removal of data from break periods serves to increase data spread, since these metrics differed more in driving periods. Excluding filter 1 (Driving and Following), filter 2 (relative velocity  30 km/h) affects the measured relative velocity the most, by Appl. Syst. Innov. 2023, 6, 55 12 of 29 removing data points corresponding to oncoming vehicles. This filter also reduced the average relative velocity to approximately 0.5 km/h for both trucks. Thus, while average relative velocities approach zero (as they should in car-following situations), they remain slightly negative. The negative value is due to most detected objects (with the exception of the preceding trucks) heading toward the sensors (as opposed to receding from them). The curated datasets for both trucks had slightly more instances of negative than positive relative velocities. The impact of filter 2 is greater for truck 3 than for truck 2, despite both having approximately equal sizes of datasets. It removes 20% of measurements from truck 3 and only 12% from truck 2, versus the number of measurements remaining after filter 1. In fact, the dataset of truck 3 enters filter 2 with a larger average relative velocity, presumably due to platoon instability and weight differences in which perturbations caused harsh braking for truck 3, which would naturally tend to occur at short distances, which filter 2 ended up removing (cf. Figure 7). Subsequent filtering steps slightly reduce the variability of relative velocity measurements, suggesting that erroneous detections are gradually removed. 3.1.2. Signal-to Noise Ratio (SNR) Filtering caused average SNR to stabilize around 29–32 dB for all trucks. Overall, filtering decreased SNR data spread. The minimum SNR was only affected by filter 3 (SNR < 15 dB). As intended, filter 3 subjected the data to a step-change, starting at 6.7–6.9 dB and ending up for all three trucks at 15.1 dB. The effects of each filter grad- ually diminish. Interestingly, all trucks measured different maximum SNR values, which were reduced in filter 5 (downsampling). As shown in Figure 5b, SNR filtering worked as intended by removing measurements at short distances. 3.1.3. Distance Distance was the only recorded metric which was not used as a basis for filtering. Average inter-vehicle distance values, shown in Table 3, suggest that truck 3 drove closer to its preceding truck than what truck 2 did. This was visually confirmed from video footage. After curation, average values were 38.6 and 36.1 m, for trucks 2 and 3, respectively. Still, distributions of distances appear to differ somewhat, with spikes at 0–10 m for truck 2, and 10–20 m for truck 3. Table 3. Distance metrics (in meters) after each filtering step. Truck Number Filtering 2 3 Step Standard Standard Average Maximum Average Maximum Deviation Deviation Raw 26.5 74.4 17.6 25.2 71.5 15.9 1 33.5 74.4 14.8 29.3 71.5 14.5 2 35.5 74.4 14.3 34.4 71.5 12.3 3 37.4 74.4 13.0 35.5 71.5 11.5 4 37.2 74.4 12.9 35.3 71.5 11.5 5 38.2 74.3 13.0 36.1 71.4 11.5 6 38.6 74.3 12.9 36.1 71.4 11.3 Unfiltered maximum distance values for trucks 2 and 3 were 74.4 and 71.5 m, respec- tively, and were virtually unaffected by filtering. Only filter 5 (downsampling) reduced maximum distances, and only by 0.1 m in both cases. This resulted in curated maximum distances of 74.3 and 71.4 m. While distance ranges were never systematically tested, maxi- mum distance values may be used as a proxy. While driving on straight road segments, inter-vehicle distances oscillated. The trucks would occasionally drive closely together before becoming dispersed, travelling with spacing between the trucks so large that each preceding truck was eventually located beyond sensor range. Thus, for each truck, measure- Appl. Syst. Innov. 2023, 6, 55 13 of 29 ments should exist at the radar range boundary. As stated, filter 6 (LWRR) did not further reduce maximum distance values than what filter 5 did. Figure 7 shows that filter 6 did not further reduce maximum distance values since it had little effect beyond approximately 50–55 m. Since LWRR video coding was carried out visually, somewhat imprecisely, the farthest radar measurements still tend to appear in LWRR-filtered data. Had it been more precise, maximum distance values following filter 5 would best represent actual upper sensor ranges. The aforementioned maximum distance values for trucks 2 and 3 fall short of the theoretical ranges 75.0 and 73.1 m by 0.9% and 2.4%, respectively. Hence, both sensors appear to underperform slightly versus expected ranges, but still fall within the estimate provided by the supplier. The deviation is smallest for truck 2. Since the datasets from both trucks are otherwise comparable, the chosen parameter values N and f for the radar s 0 sensor in truck 2 appears to be preferable. Future testing could explore this. For the curated radar data, average relative velocity and average SNR were inspected as a function of inter-vehicle distances. For both trucks 2 and 3, these metrics were calcu- lated within successive 10-m distance bins. The lower bin was 0–10 m, and the upper bin was 70 m and above. For all bins, average relative velocities fell within the error margin of the sensors. On the other hand, average SNR values are more interesting. For all bins, average SNR very seldomly fell below 30 dB, and remained high even at long distances, indicating that the radar cross-sections of the preceding trucks are sufficiently large to allow for longer detection ranges, cf. Tables A8 and A9 in Appendix A.4. Thus, parameters could likely have been chosen to achieve a maximum range approaching, and perhaps even exceeding, 100 m. The radar user manual illustrates that cars can be detected at 75 m, while buildings can be detected at 100 m. Due to the size of the truck rear walls, they may provide ‘building-like’ detection ranges. This study tracked large, reflective metal back walls of semi-trailers, with cross-sectional areas exceeding 8 square meters. The back walls of semi-trailers are larger than those of other vehicles and presumably facilitate higher-quality detections. Measuring distances to smaller vehicles should be tested. While the maximum distance range may be greater when detecting trucks than passenger cars, situations may exist where trucks are less favorable. For instance, when traversing sharp curves at short distances, the radar cross-section changes as truck back walls change angles in relation to the sensors, as opposed to being located perpendicular to them. Some military ships and aircraft are deliberately made from planar surfaces joined at sharp angles to achieve radar stealth. Similarly, truck back walls are also two-dimensional planar surfaces. In curves, reflected signals may scatter away from the receivers and cause data loss. This may also occur if the preceding truck is located outside the main lobe of the radar antenna. Such diffraction effects were reported in [38]. 3.2. Radar Sensor Operation in Different Driving Situations The curated radar data were coupled with video codes to determine whether the sensors captured high-quality inter-vehicle distance measurements when they should have been able to. Filter 6 aimed to remove periods when respective preceding trucks were located beyond radar range. We now check whether the radars were able to account for the remaining duration. Driving is explored, and also Tunnel and Roundabout subcategories. Outputs from filter 6 contain radar data at 1 Hz, so aggregated video code durations (in seconds), when filtered for LRWW, are directly comparable to curated radar data. The proportions are shown in Table 4. Both sensors retained a similar number of measurements in each condition. As Table 4 shows, the sensors detected the preceding truck in most situations where it had the opportunity to do so. Using Tunnels to illustrate: Trucks 2 and 3 drove in Tunnels while having their respective preceding trucks within radar range (LWRR) for approximately 26 min (1647 and 1688 s, respectively). Aggregated over these same periods, their radar sensors outputted 1576 and 1609 curated measurements (at 1 Hz) after filter 6. As proportions, this yields 96% and 95%, respectively. Examples from each condition are shown in Table A5 in Appendix A.2. Appl. Syst. Innov. 2023, 6, 55 14 of 29 Table 4. Proportions of curated radar data retained vs. LWRR-filtered aggregated event durations. Trucks Condition 2 3 Driving * 85% 83% Tunnels 96% 95% Roundabouts 88% 89% * Driving includes Tunnels and Roundabouts. Radar data and video codes can be used to visualize and explore excerpts of curated radar data (LWRR) from different driving intervals. Inter-vehicle distances are plotted versus time. Excerpts indicate large variability in inter-vehicle distances as a function of infrastructure type and road standard, as suggested in [23]. Since Tunnel and Roundabout video codes have non-zero durations, a choice had to be made regarding from which truck to visualize them. Truck 2 was chosen, being the middle truck in the platoon for 88% of the drive. Note that, at the time resolution used (1 min divisions), the difference would have been negligible if visualizing codes from one of the other two vehicles. Oncoming trucks seemed to cause the platoon to slow down, which also affected inter-vehicle distances, particularly on narrow roads and in sharp curves. Therefore, Oncoming truck video codes are shown. Being point events, these could be visualized for each truck separately. The horizontal axes (d hh:mm) were not fixed, so excerpts have slightly different durations (between 5 to 10 min). As the field study took place over two days, d-values are either 1 or 2. 3.2.1. Visual Verification of Maximum Range Figure 8 shows three excerpts from Driving. While no tunnels or roundabouts are shown, statistics from this category in Table 4 also include those durations, thus showcasing diverse, complex driving segments. While serving as followers for 7 h, trucks 2 and 3 drove with their respective preceding trucks within radar range for approximately 6 h. All excerpts in Figure 8 had the preceding truck within range, so trends for measured distances appear mostly continuous. The top excerpt stems from an old, narrow road section without centerlines. Trucks often adjusted their speeds, including when encountering opposing trucks, as revealed by reduced inter-vehicle distances. The middle excerpt stems from the traversal of a flat and wide high-quality road with a 90 km/h speed limit and gentle horizontal curves. Here, the trucks maintained constant distances over long time periods, and opposing trucks did not influence the platoon. The individual data points from truck 2 which are located below the general trend were chosen by the algorithm since the sensor did not detect other data points during that logging instance. In general, truck 2 had more such erroneous measurements than truck 3 did, but it is unclear why this is the case. The lower excerpt illustrates the descent of a challenging mountain pass (negative 6% gradient), where trucks 2 and 3 reduced their speed repeatedly, to avoid speeding and becoming located too close to their respective preceding truck. These excerpts suggest that truck platooning is more suitable on wide, modern roads than on old roads with adverse horizontal and vertical alignment. Appl. Syst. Innov. 2023, 6, x FOR PEER REVIEW 15 of 29 Appl. Syst. Innov. 2023, 6, 55 15 of 29 that truck platooning is more suitable on wide, modern roads than on old roads with ad- verse horizontal and vertical alignment. Figure 8. Inter-vehicle distances measured from three separate driving excerpts, LWRR, during Figure 8. Inter-vehicle distances measured from three separate driving excerpts, LWRR, during which which radar range was never exceeded. radar range was never exceeded. Figure 9 shows three excerpts during which, for both trucks 2 and 3, the inter-vehicle Figure 9 shows three excerpts during which, for both trucks 2 and 3, the inter-vehicle distance appears to have occasionally exceeded maximum radar ranges. Blue (truck 2) and distance appears to have occasionally exceeded maximum radar ranges. Blue (truck 2) green (truck 3) horizontal lines illustrate the farthest distances detected, which differ and green (truck 3) horizontal lines illustrate the farthest distances detected, which differ somewhat somewhat between between exce excerpts. rpts. Nonetheless, Nonetheless, for forradar radarsensors sensorsin intr tr ucks ucks2 2and and3, 3, re respectively spectively, , the the m maximum aximum distances distances ar eare appr ap oximately proximately 75 and 75 and 70 m, 70 m, wh which support ich support the the aforementioned aforemen- statistics-based tioned statistics-based r radar range adestimates. ar range estim The thr ates. The three excerpts stem f ee excerpts stem from two dif rom two di ferent mountain fferent passes. mounta The in passes. The two upper excerpts corre two upper excerpts correspond to traversal spond to tra of Mountain versal ofpass Mount A, a with in pa very ss A, dif with v ficultecombinations ry difficult com ofbsharp inations o horizontal f sharp horizont curves and al c vertical urves and v gradient ertical s. V gr ideos adient wer s. V e ialso deos useful in inspecting the radar data after curation. The combination of long inter-vehicle were also useful in inspecting the radar data after curation. The combination of long inter- distances and horizontal curves occasionally caused the preceding truck to be obscured by vehicle distances and horizontal curves occasionally caused the preceding truck to be ob- rock walls at the inner part of right-turn curves (pale shading). Mountain pass B, shown in scured by rock walls at the inner part of right-turn curves (pale shading). Mountain pass the lowermost excerpt, was more forgiving in terms of road alignment. B, shown in the lowermost excerpt, was more forgiving in terms of road alignment. In the upper excerpt, the data from truck 3 (green) at timestamp 16:56 potentially reveals the presence of a phenomenon which, together with the high average SNR values measured at large distances, suggests that sensors are capable of measuring the preceding truck far away. It appears as if measurements which naturally belong to the top of the green curve are folded down, instead of occurring at 90–100 m, where extrapolation would place them. Thus, it looks like the radar in truck 3 does measure the preceding truck, despite it being located beyond the maximum range imposed by the chosen parameters. Stated otherwise, the radar appears to measure points beyond its unambiguous range, referring to the maximum distance a target can have while it can be guaranteed that the reflected pulse from that target corresponds to the most recent transmitted pulse [14,39]. At this timestamp, it appears as if the returned signal is associated with the wrong transmitted pulse, so Appl. Syst. Innov. 2023, 6, 55 16 of 29 the range becomes ambiguous. Filtering for LWRR and LBRR served to remove most durations where such long-distance samples may have been folded into the ambiguous Appl. Syst. Innov. 2023, 6, x FOR PEER REVIEW 16 of 29 range. However, the manual video coding process did not remove all such instances. The presence of folding may have influenced distance metrics to the downside. Figure 9. Inter-vehicle distances measured from three separate driving excerpts, LWRR, during Figure 9. Inter-vehicle distances measured from three separate driving excerpts, LWRR, during which which radar range was exceeded. Horizontal lines show farthest measurements. radar range was exceeded. Horizontal lines show farthest measurements. In the upper Data from truck excerpt, the 2 in the da middle ta from excerpt truck 3 (green) of Figure 9 atar tiemest noisier amp 16 than :56 potentia for all other lly reveals the presence of a phenomenon which, together with the high average SNR values excerpts shown, but it is unclear whether the folding phenomenon occurs here. The outage measured at large distances, suggests that sensors are capable of measuring the preceding for truck 2 in the lower excerpt stems from a period when a passenger car partly overtook truck the platoon, far away. It being a sandwiched ppears as ifbetween measurements whi truck 1 andc2h until natura also lly belong to the top of overtaking the lead tr the uck. green c Data from urvsuch e areperiods folded d wer own, in e removed steadby of occur filter 1 r (Driving ing at 90–100 m and Following , where ). extrapolation would place them. Thus, it looks like the radar in truck 3 does measure the preceding 3.2.2. Tunnels truck, despite it being located beyond the maximum range imposed by the chosen param- eters. St It isated unclear otherwise, th how tunnels e radaf ar fect appears to measure points bey the ability of the sensors to o measur nd its un e inter ambiguo -vehicle us distances. Tunnels may reduce operational complexity, as rock walls cause peripheral range, referring to the maximum distance a target can have while it can be guaranteed narrowing of roadside areas. However, walls, lighting and ventilation elements may that the reflected pulse from that target corresponds to the most recent transmitt ed pulse introduce clutter. Such features are less frequent on roads in natural terrain. [14,39]. At this timestamp, it appears as if the returned signal is associated with the wrong All Tunnel driving occurred in truck order 1–2–3. Figure 10 shows excerpts from transmitt ed pulse, so the range becomes ambiguous. Filtering for LWRR and LBRR served six representative Tunnel traversals (green shading). Driving periods outside tunnels have to remove most durations where such long-distance samples may have been folded into white backdrops. All excerpts were coded as LWRR, except for the period between the the ambiguous range. However, the manual video coding process did not remove all such two tunnels in the top excerpt, when data were lacking for truck 2. This period was instances. The presence of folding may have influenced distance metrics to the downside. coded as LBRR, as truck 1 was located far away. Comparing Figures 8 and 10, it seems Data from truck 2 in the middle excerpt of Figure 9 are noisier than for all other ex- as if filter 4 (target selection) chooses erroneous data points at comparable frequencies cerpts shown, but it is unclear whether the folding phenomenon occurs here. The outage both inside and outside tunnels. Thus, tunnel driving does not appear to degrade radar for truck 2 in the lower excerpt stems from a period when a passenger car partly overtook operating conditions. In tunnels, maximum distance values for trucks 2 and 3 were reduced the platoon, being sandwiched between truck 1 and 2 until also overtaking the lead truck. by 8–9%, and the distance standard deviation dropped by 18% and 15%, respectively. Data from such periods were removed by filter 1 (Driving and Following). Thus, inter-vehicle distances were moderated by Tunnels, causing closer, more uniform driving at lower speeds. This made preceding trucks occupy a larger part of radar the field- 3.2.2. Tunnels It is unclear how tunnels affect the ability of the sensors to measure inter-vehicle dis- tances. Tunnels may reduce operational complexity, as rock walls cause peripheral nar- rowing of roadside areas. However, walls, lighting and ventilation elements may intro- duce clutt er. Such features are less frequent on roads in natural terrain. Appl. Syst. Innov. 2023, 6, x FOR PEER REVIEW 17 of 29 All Tunnel driving occurred in truck order 1–2–3. Figure 10 shows excerpts from six representative Tunnel traversals (green shading). Driving periods outside tunnels have white backdrops. All excerpts were coded as LWRR, except for the period between the two tunnels in the top excerpt, when data were lacking for truck 2. This period was coded as LBRR, as truck 1 was located far away. Comparing Figures 10 and 8, it seems as if filter 4 (target selection) chooses erroneous data points at comparable frequencies both inside and outside tunnels. Thus, tunnel driving does not appear to degrade radar operating condi- tions. In tunnels, maximum distance values for trucks 2 and 3 were reduced by 8–9%, and the distance standard deviation dropped by 18% and 15%, respectively. Thus, inter-vehi- Appl. Syst. Innov. 2023, 6, 55 17 of 29 cle distances were moderated by Tunnels, causing closer, more uniform driving at lower speeds. This made preceding trucks occupy a larger part of radar the field-of-view. How- ever, measurements collected in tunnels were noisier than those collected during Driving, of-view. However, measurements collected in tunnels were noisier than those collected indicated by lower SNR values (weaker signal). For both followers, Tunnel filtering re- during Driving, indicated by lower SNR values (weaker signal). For both followers, Tunnel duced mean SNR by 3%, maximum values by 6–8% and standard deviations by 7–11%. filtering reduced mean SNR by 3%, maximum values by 6–8% and standard deviations by This did not affect the curated radar data when plott ed: In Figure 9, erroneously selected 7–11%. This did not affect the curated radar data when plotted: In Figure 9, erroneously targets in tunnels appear to have distance values at similar deviations to the trend, com- selected targets in tunnels appear to have distance values at similar deviations to the trend, pared to erroneously selected targets outside tunnels. Thus, it appears as if inter-vehicle compared to erroneously selected targets outside tunnels. Thus, it appears as if inter-vehicle distance measurements between platooning trucks are not adversely affected by tunnels. distance measurements between platooning trucks are not adversely affected by tunnels. Figure 10. Radar data for three separate truck driving excerpts, Tunnel and LWRR. Figure 10. Radar data for three separate truck driving excerpts, Tunnel and LWRR. 3. 3.2.3. 2.3. Ro Roundabouts undabouts Roundabouts may allow for exploring the effects of sharp road curvature on radar Roundabouts may allow for exploring the effects of sharp road curvature on radar operation. Most roundabout traversals involved straight movements, encountering little to operation. Most roundabout traversals involved straight movements, encountering litt le no traffic. Figure 11 shows excerpts from five traversals (pale red shading), all of which to no traffic. Figure 11 shows excerpts from five traversals (pale red shading), all of which occurred in truck order 1–2–3. Driving periods outside roundabouts have white backdrops. occurred in truck order 1–2–3. Driving periods outside roundabouts have white back- The first traversal in the middle excerpt involved trucks 1 and 2 performing a full revolution drops. The first traversal in the middle excerpt involved trucks 1 and 2 performing a full to get rid of external vehicles located between trucks 2 and 3. Truck 3 had Other vehicles revolution to get rid of external vehicles located between trucks 2 and 3. Truck 3 had Other preceding it before entering this roundabout, so its data were removed by filter 1. The vehicles preceding it before entering this roundabout, so its data were removed by filter 1. traversal in the lower excerpt involved all three trucks revolving one round. Trucks 2 and The traversal in the lower excerpt involved all three trucks revolving one round. Trucks 2 3 both had 10 traversals as followers, with 5 straight, 1 left turn and 4 right turns. For and 3 both had 10 traversals as followers, with 5 straight, 1 left turn and 4 right turns. For both trucks, only half of the aggregated Roundabout durations had the preceding truck within field-of-view, and even when accounting for field-of-view, 11–12% of data are lost. Figure 11 shows that data points retained in roundabouts generally have scattered distance values which are too large to represent the preceding truck. Sharp curve radii and limited antenna beam width resulted in lost field-of-view to the preceding truck, so radar sensors detected irrelevant objects until field-of-view was regained after the turn. Appl. Syst. Innov. 2023, 6, x FOR PEER REVIEW 18 of 29 both trucks, only half of the aggregated Roundabout durations had the preceding truck within field-of-view, and even when accounting for field-of-view, 11–12% of data are lost. Figure 11 shows that data points retained in roundabouts generally have scatt ered dis- tance values which are too large to represent the preceding truck. Sharp curve radii and Appl. Syst. Innov. 2023, 6, 55 18 of 29 limited antenna beam width resulted in lost field-of-view to the preceding truck, so radar sensors detected irrelevant objects until field-of-view was regained after the turn. Figure 11. Radar data from three separate driving excerpts with Roundabouts, LWRR. Figure 11. Radar data from three separate driving excerpts with Roundabouts, LWRR. Average SNR values were 30 and 32% lower for trucks 2 and 3 in Roundabouts than Average SNR values were 30 and 32% lower for trucks 2 and 3 in Roundabouts than for for Driving, respectively, so the radar sensors detected noisier data. Proportions of radar Driving, respectively, so the radar sensors detected noisier data. Proportions of radar data data as a function of LWRR-filtered Roundabout durations (88–89%), are greater than for as a function of LWRR-filtered Roundabout durations (88–89%), are greater than for Driving Driving (83–85%). This is likely due to the trucks being grouped closer together when trav- (83–85%). This is likely due to the trucks being grouped closer together when traversing ersing Roundabouts. Speeds were also lower, causing smaller inter-vehicle distances and Roundabouts. Speeds were also lower, causing smaller inter-vehicle distances and greater greater spatial concentration of measurements. For trucks 2 and 3, mean distances were spatial concentration of measurements. For trucks 2 and 3, mean distances were 66–62% 66–62% shorter during Roundabouts than during Driving. The exploration shows that pre- shorter during Roundabouts than during Driving. The exploration shows that preceding ceding trucks are tracked poorly when the platoon passes through roundabouts. trucks are tracked poorly when the platoon passes through roundabouts. 3.3. Suggestions for Future Work 3.3. Suggestions for Future Work This study explored the extent to which uRAD radar sensors could capture inter- This study explored the extent to which uRAD radar sensors could capture inter- vehicle distances during truck platooning. Several suggestions have been identified. vehicle distances during truck platooning. Several suggestions have been identified. First, mode 4 (7 Hz) might be more suitable than mode 3 (13 Hz). As data were First, mode 4 (7 Hz) might be more suitable than mode 3 (13 Hz). As data were downsampled anyway, the trade-off between higher update rate (mode 3) and enhanced downsampled anyway, the trade-off between higher update rate (mode 3) and enhanced properties for ghost target reductions in complex scenarios (mode 4) should be explored. properties for ghost target reductions in complex scenarios (mode 4) should be explored. To simplify data collection, radars could be remotely engaged from escort vehicles. If To simplify data collection, radars could be remotely engaged from escort vehicles. field trials are undertaken in areas with adequate cell coverage, virtual network computing If field trials are undertaken in areas with adequate cell coverage, virtual network compu- (VNC) could be used to remote control all Raspberry Pi microprocessors. Radar data may ting (VNC) could be used to remote control all Raspberry Pi microprocessors. Radar data also be visualized and coded in real-time. Future work may validate the radar data against may also be visualized and coded in real-time. Future work may validate the radar data GNSS positions, if undertaken in areas with good conditions for GNSS receivers. The against GNSS positions, if undertaken in areas with good conditions for GNSS receivers. performance of the sensors should also be compared against a known baseline, i.e., radars The performance of the sensors should also be compared against a known baseline, i.e., with known characteristics, such as those listed in [14], in controlled environments. radars with known characteristics, such as those listed in [14], in controlled environments. The cross-sectional signature of the trucks was not measured, and scattering effects were not explored. If present, such effects would be reflected in SNR values if systematically inspected at equal distances while varying the angle of the back wall of the preceding truck, resembling the set-up in [40]. Windshield attenuation effects were also not studied. Synchronization of videos and radar data, and the subsequent process of video coding, worked well. However, both should preferably be automated, to reduce post-processing Appl. Syst. Innov. 2023, 6, 55 19 of 29 efforts and related human errors. Traffic and infrastructure events may be identified directly from radar data. Herein, tunnels and roundabouts were coded from the moment where each respective truck entered them, and to the moment when each respective truck left them. Later, video codes were overlaid on the radar data collected from each truck. However, since radar data shows the preceding truck, this may have introduced a systematic error. Perhaps video codes from the preceding truck should have been used instead. For tunnels, this is not particularly problematic since tunnel traversals (with LWRR) had long durations (on average 1.1 min). Thus, entering and leaving the tunnel occupies a very small part of the 1 1 total duration. For roundabouts, however, the preceding truck had often traversed – of 4 2 the roundabout before the truck in question entered it, and it was first coded. Distinctions may be made between the preceding truck being located beyond range, and it being located laterally beyond field-of-view, as these are different phenomena. All data points may also be given metadata for all relevant video codes, simplifying video inspections of interesting events in the data. Statistical approaches may allow cut-off values for filters 2 and 3 to be chosen auto- matically. In the target detection step, distance filtering could also be considered, perhaps discarding data points with distances deviating significantly from the general trend. This may solve the problem of erroneous single-object detections. Established data filtering, target tracking [41] or clustering techniques [42–44] could also be used, alongside more computationally complex methods for annotating or labelling combinations of radar and camera footage, e.g., in [31,45], and perhaps also machine learning approaches [46]. At times, dashboard cameras malfunctioned due to power issues, totaling 6% of driving time, during which 10% of all radar data were logged. As it was not possible to de- termine whether the sensors had reasonable operating conditions in these periods, the data were discarded. Mitigations include redundant cameras and independent power supplies. 4. Conclusions Anteral uRAD radar sensors for Raspberry Pi were tested for estimating inter-vehicle distances between trucks. Three trucks participated in a real-word platooning field study. Data from integrated sensors were unavailable. Comparable results were found from the sensors in the two rearmost trucks, suggesting that they are feasible for this use case. Data filtering involved a multi-faceted methodology. While also filtering based on relative velocity and signal-to-noise ratio, video footage allowed for removal of data from irrelevant periods, and for exploring sensor operation in roundabouts and tunnels. This would not have been possible without video footage. The curated radar data can be used to model expected fuel savings from truck platooning on specific types of roads and road features. Sensor ranges were estimated at 74 and 71 m, i.e., slightly shorter than suggested by theoretical calculations. The sensors captured the preceding truck for 83–85% of the time when it was located within radar range. In tunnels specifically, 95–96% of driving time was accounted for, likely due to closer driving. Average SNR decreased 3% in tunnels, compared to all driving, but this did not appear to affect the accuracy of the target detection step. When turning in roundabouts, the field-of-view to the preceding truck was often lost, and the sensors detected their surroundings until field-of-view was regained after completing the turn, causing average SNR values to drop (30–32% lower). Straight movements in roundabouts were less problematic, as field-of-view was mostly retained. High SNR values were observed at far distances, indicating that the sensors, with optimal parameters, may be capable of measuring preceding trucks further away. The findings suggest that simple, inexpensive radar sensors and action cameras can facilitate collection of inter-vehicle distance data from truck platooning field trials. Appl. Syst. Innov. 2023, 6, x FOR PEER REVIEW 20 of 29 findings suggest that simple, inexpensive radar sensors and action cameras can facilitate collection of inter-vehicle distance data from truck platooning field trials. Appl. Syst. Innov. 2023, 6, 55 20 of 29 Author Contributions: Conceptualization, M.M.L.; methodology, M.M.L., T.T. (Thomas Thoresen), M.H.R.E.; software, T.T. (Thomas Thoresen); validation, T.T. (Thomas Thoresen); formal analysis, M.M.L.; investigation, M.M.L.; data curation, T.T. (Thomas Thoresen); writing—original draft prep- aration, M.M.L.; writing—review and editing, M.M.L., M.H.R.E., T.T. (Trude Tørset), T.L.; visuali- Author Contributions: Conceptualization, M.M.L.; methodology, M.M.L., T.T. (Thomas Thoresen), zation, T.T. (Thomas Thoresen); supervision, T.T. (Trude Tørset), T.L.; project management, M.M.L.; M.H.R.E.; software, T.T. (Thomas Thoresen); validation, T.T. (Thomas Thoresen); formal analysis, funding acquisition, T.L. All authors have read and agreed to the published version of the manu- M.M.L.; investigation, M.M.L.; data curation, T.T. (Thomas Thoresen); writing—original draft prepa- script. ration, M.M.L.; writing—review and editing, M.M.L., M.H.R.E., T.T. (Trude Tørset), T.L.; visualization, T.T. (Thomas Thoresen); supervision, T.T. (Trude Tørset), T.L.; project management, M.M.L.; funding Funding: This research was funded in part by the Norwegian University of Science and Technology acquisition, T.L. All authors have read and agreed to the published version of the manuscript. (NTNU) and the Norwegian Public Roads Administration (NPRA), through the Innovation and Im- plementation initiative (D11351). The field study itself, titled “Smart and Connected Truck Train” Funding: This research was funded in part by the Norwegian University of Science and Technology was financed by the NPRA, cf. the Norwegian database for public procurement. Notice of intent for (NTNU) and the Norwegian Public Roads Administration (NPRA), through the Innovation and procurement of transportation research services. “Smart and Connected Truck Train”. Norwegian Implementation initiative (D11351). The field study itself, titled “Smart and Connected Truck Train” national database for public procurement: htt ps://doffin.no/nn/Notice/Details/2020-301190 (ac- was financed by the NPRA, cf. the Norwegian database for public procurement: https://doffin.no/ cessed 3 May 2023). nn/Notice/Details/2020-301190. Institutional Review Board Statement: Not applicable. Institutional Review Board Statement: Not applicable. Informed Consent Statement: Informed consent was obtained from all participants involved. Informed Consent Statement: Informed consent was obtained from all participants involved. Data Availability Statement: Data are available on request from the corresponding author. Data Availability Statement: Data are available on request from the corresponding author. Acknowledgments: The main author would like to acknowledge the Anteral supplier team for an- Acknowledgments: The main author would like to acknowledge the Anteral supplier team for swering inquiries. answering inquiries. Conflicts of Interest: The authors declare no conflict of interest. Conflicts of Interest: The authors declare no conflict of interest. Appendix A Appendix A The appendix is subdivided into four parts, providing background information and The appendix is subdivided into four parts, providing background information and methodological details for reproducibility, alongside tables with results. methodological details for reproducibility, alongside tables with results. Appendix A.1. The Radar Sensors Appendix A.1. The Radar Sensors Anteral uRAD radar sensors, version 1.1, were used alongside Raspberry Pi 4, model Anteral uRAD radar sensors, version 1.1, were used alongside Raspberry Pi 4, model B B with 4 GB RAM. Micro-SD cards (16 GB) were purchased, with the Raspberry Pi oper- with 4 GB RAM. Micro-SD cards (16 GB) were purchased, with the Raspberry Pi operating ating system and uRAD software pre-installed. Technical support was also purchased. Ta- system and uRAD software pre-installed. Technical support was also purchased. Table A1 ble A1 provides an overview of the radar configuration parameters. Testing prior to the provides an overview of the radar configuration parameters. Testing prior to the field trial is field trial is outlined in Figure A1 and Table A2. Collected test data and corresponding outlined in Figure A1 and Table A2. Collected test data and corresponding camera footage camera footage were sent to the supplier, who proposed recommendations for future test- were sent to the supplier, who proposed recommendations for future testing. Table A3 ing. Table A3 details the parameters which were used in the field study. details the parameters which were used in the field study. (a) (b) (c) Figure A1. (a) Dump truck test set-up; (b) Back wall of preceding truck seen from cabin during Figure A1. (a) Dump truck test set-up; (b) Back wall of preceding truck seen from cabin during driving, with radar sensor (yellow) and dashcam (red) circled; (c) Radar sensor side view. driving, with radar sensor (yellow) and dashcam (red) circled; (c) Radar sensor side view. The supplier detailed the pre-processing steps taken before sensors store data to mem- ory: The radar transceiver chip receives the reflected signal. The mixer mixes the received (RX) signal with the transmitted (TX) signal, and outputs in-phase (I) and quadrature (Q) components at Intermediate Frequency (IF). These two analog I/Q IF signals go through a low-pass filter, an amplifier stage and a high-pass filter. Filter values are proprietary. Subse- quently, the analog signal is digitalized with an ADC at 25 kHz in mode 1 and 200 kHz in Appl. Syst. Innov. 2023, 6, 55 21 of 29 modes 2, 3 and 4. In the digital domain, the complex signal is formed, FFT obtained, and from it, range, velocity and SNR of detected targets are derived. The supplier also stated that calibration by the user is not needed, as performance is controlled in the lab prior to shipment, by measuring a constant distance of 1.5 m in mode 2. The combination of one Raspberry Pi attached to one uRAD radar board is here termed a device. The device was fixed to the interior of the windshield using a universal phone suction mount with a flexible arm, oriented such that the USB-C and micro-HDMI ports pointed directly upwards (cf. Figure A1). Since Raspberry Pi microprocessors were powered on and off using a USB-C cable, this orientation facilitated easy access and line of sight from above, for cable insertion and removal. It also made sure that the sensor did not detect the mount itself, being located outside radar field-of-view. Powerbanks and USB-C cables powered the devices at optimum voltage and amperage, while making data logging independent of truck power systems and status. This eliminated potential issues with undervolting and voltage spikes from in-vehicle outlets, while leaving flexibility for when logging start and stop had to be administered. Each Raspberry Pi had a Bluetooth USB dongle for a wireless keyboard and mouse, minimizing direct device contact. The dongle added an additional reason for ensuring stable power supply. The radar supplier confirmed that the dongle radio frequency would not affect radar operation. Along with the dongle, the micro-HDMI and power cables remained plugged into the device throughout the field trial. This allowed the devices to remain vehicle-mounted throughout both days, requiring only insertion and removal of the far-ends of the cables into the screen and powerbank for interfacing with the devices and powering them devices on and off, respectively. Radar output files never exceeded 5 MB, i.e., they were unproblematic with respect to SD card storage capacity. A battery powered portable monitor was used. With 30 vertical fields-of-view, 15 swept down from horizontal, so vertical sweep became 75 , assuming sensors were mounted perfectly level. If trucks, with uRAD antennas at height h , traversed a constant gradient road section, the road would be detected at a distance, d, given Equation (A1): d = h tan(75 ) (A1) The presence of aftermarket dashboard tabletop surfaces required placing sensors at slightly different heights in each truck during the field study. With radar antennas at heights of 2.29 m (truck 2) and 2.15 m (truck 3), the road would be detected 8–8.5 m forwards. Having 30 fields-of-view also in the horizontal direction, radars also saw this far sideways at road level. Table A1. Overview and discussion of uRAD radar parameters. Parameter Discussion Of four modes available, only modes 3 (triangular) and 4 (dual-rate) measured both distance and velocity. Velocity would enable filtering away stationary and oncoming objects, to be left with desired inter-vehicle distance to the preceding truck. Modes 3 and 4 differed in upper distance range and update rate. Mode 3 had an upper distance range of 100 m, versus 75 m for mode 4. The supplier stated Mode that the range would also depend on the target, meaning its radar cross-section: “(. . . ) a person is detected up to 40 m. (. . . ) a truck, that is bigger and reflects more, (. . . ) will be detected [at] 70 m but probably (. . . ) much farther. 100 m is not a limitation of the radar, [but] a guide (. . . ) for very big targets.” Mode 4 should reduce ghost target detections in multi-target scenarios, at the expense of reduced range. Appl. Syst. Innov. 2023, 6, 55 22 of 29 Appl. Syst. Innov. 2023, 6, x FOR PEER REVIEW 22 of 29 Table A1. Cont. Parameter Discussion Ramp start frequency, f , could be set as 5–195 for modes 2–4. Operation Ramp start frequency, f0, could be set as 5–195 for modes 2–4. Operation bandwidth, BW, meaning bandwidth, BW, meaning the frequency sweep used in modes 2–4, depends on f , the frequency sweep used in modes 2–4, depends on f0, and should be maximized, subject to and should be maximized, subject to Equation (A2), to increase accuracy and to Equation (A2), to increase accuracy and to distinguish closely located targets. For each radar, dif- distinguish closely located targets. For each radar, different values were chosen to Ramp start freq., f0. Ramp start ferent v freq., f a . lues were chosen to avoid interference. avoid interference. Operation band- Operation bandwidth, BW 𝐵𝑊 = 245𝑓 BW = 245 f (A2) (A2) max width, BW The f0 parameter denotes the starting frequency of the waves emitt ed by the sensor. The sensor op- The f parameter denotes the starting frequency of the waves emitted by the sensor. The sensor operates at a frequency bandwidth of 24.005–24.245 GHz, and erates at a frequency bandwidth of 24.005–24.245 GHz, and f0 values are defined (in MHz) to set f values are defined (in MHz) to set the offset from the lower threshold. the offset from the lower threshold N is the number of samples taken from the reflected wave to calculate distance Ns is the number of samp s les taken from the reflected wave to calculate distance and velocity. and velocity. Highest update rate requires lowest possible N . However, a trade-off Highest update rate requires lowest possible Ns. However, a trade-off is needed, since BW and Ns is needed, since BW and N determine maximum range, through Equation (A3). determine maximum range, through Equation (A3). Samples and ramp Ns 𝑐𝑒𝐷𝑖𝑠𝑡𝑎𝑛 =75∙ (A3) Samples and ramp duration, N s Distance = 75 (A3) max duration, Ns BW The Ns parameter serves two purposes. Firstly, it defines the duration of each wave ramp, and sec- The N parameter serves two purposes. Firstly, it defines the duration of each ondly, it outlines the sampling rate from the reflected wave, per ramp duration, which can be wave ramp, and secondly, it outlines the sampling rate from the reflected wave, used to calculate output metrics. per ramp duration, which can be used to calculate output metrics. Ntar is the number of targets that the sensor detects, 5 being maximum. If detecting more objects, Max. detected tar- N is the number of targets that the sensor detects, 5 being maximum. If detecting tar the sensor logs data for those 5 with highest SNR. Ntar was maximized, capturing most data and more objects, the sensor logs data for those 5 with highest SNR. N was tar gets, Ntar Max. detected targets, N tar providing possibility for filtering unwanted objects later. maximized, capturing most data and providing possibility for filtering unwanted objects later. For modes 2–4, Rmax is the maximum distance below which targets will be detected. Rmax artificially reduces the zone of interest, excluding For modes 2–4, R targets beyo is the maximum nd this distance distanc below e, evwhich en if they targets hav will e hbe igher SNR max detected. R artificially reduces the zone of interest, excluding targets beyond than those within it. Rmax was cho max sen as 100 for all sensors, as this would search targets within the this distance, even if they have higher SNR than those within it. R was chosen max Maximum detec- entire range. When asked if the sensors would stay fixed on the preceding truck in horizontal as 100 for all sensors, as this would search targets within the entire range. When tion distance, Rmax curves, the supplier stated that manual antenna modification could double the horizontal FOV, to asked if the sensors would stay fixed on the preceding truck in horizontal curves, Maximum detection distance, R max the detriment of upper detection range. No manual modifications were made. For vertical curves, the supplier stated that manual antenna modification could double the horizontal FOV, to the detriment of upper detection range. No manual modifications were the supplier cited that the road in front of the truck, which would be more visible in vertical sag made. For vertical curves, the supplier cited that the road in front of the truck, curves, could reflect the signal, masking the preceding truck. which would be more visible in vertical sag curves, could reflect the signal, Moving target indi- masking the preceding truck. cator, MTI. Moving target indicator (MTI) allowed for including data only from objects with motion relative Moving target indicator (MTI) allowed for including data only from objects with Moving target indi-cator, MTI. Movement detec- to the sensor. Mth is only relevant when using uRAD as a movement detector, and was not used. motion relative to the sensor. Mth is only relevant when using uRAD as a Movement detection, Mth tion, Mth movement detector, and was not used. Table A2. Pre-trial testing of radar parameter configurations. Table A2. Pre-trial testing of radar parameter configurations. Supplier Modifications and Recommen- Pre-Test Steps Parameters User Experience Supplier Modifications and dations Pre-Test Steps Parameters User Experience Recommendations • Switch off Mth, as it is not relevant for the application. Switch off Mth, as it is not • Preceding traffic recorded well. relevant for the application. • Replaced GUI with Python script for Preceding traffic recorded well. 1: Passenger car Mode = 2 Replaced GUI with Python script • Data writt en to the same file each time Data written to the same file each increased update rate. test with one radar for increased update rate. f0 = 45 Mode = 2 1: Passenger car test subsequent data collection is stopped time subsequent data collection is • Code rewri Code rewritten tt en to crea to create te new output new sensor and stand- f = 45 with one radar sensor BW = 20 0 0 and started. stopped and started. output files upon each logging files upon each logging start. BW = 200 and standard ard graphical user Data are only written to file upon Ns = 200 • Data are only writt en to file upon log- start. N = 200 graphical user logging stop. Susceptible to data • Output files are now named with interface (GUI) ging stop. Susceptible to data loss if Ntar = 5 Output files are now named with N = 5 interface (GUI) tar loss if equipment malfunctions. “start logging time” in human- reada- equipment malfunctions. Cannot dis- “start logging time” in human- Rmax = 100 R = 100 max Cannot distinguish driving ble form readable at. format. tinguish driving segments. MTI = 1 MTI = 1 segments. Data are now continuously Mth = 1• Epoch time format impractical and • Data are now continuously writt en to Mth = 1  Epoch time format impractical written to file during logging, as not human-readable. and not human-readable. file during logging, as opposed to opposed to batch writing upon batch writing upon logging termina- logging termination. tion. 2: Test with two Mode = 2, 3 • The preceding truck was recorded • Field study involves three sensors, dump trucks. Fol- f0 = 45 well, except in curves and intersec- not one. lower with BW = 200 tions. Appl. Syst. Innov. 2023, 6, 55 23 of 29 Table A2. Cont. Supplier Modifications and Pre-Test Steps Parameters User Experience Recommendations Appl. Syst. Innov. 2023, 6, x FOR PEER REVIEW 23 of 29 The preceding truck was recorded well, except in curves and Mode = 2, 3 2: Test with two dump intersections.  Field study involves three sensors, f = 45 trucks. Follower with Mode 3 is preferred over mode 2; not one. dashcam and one Ns = 200 BW = 200 • Mode 3 is preferred over mode 2;• Finalized parameters were recom- dashcam and one makes it easy to filter away  Finalized parameters were N = 200 radar sensor Ntar = 5 makes it easy to filter away stationary mended to avoid interference, yet radar sensor stationary objects and oncoming recommended to avoid N = 5 tar Rmax = 100 objects and oncoming vehicles based maximize sensor range. vehicles based on relative velocity. interference, yet maximize sensor R = 100 max MTI = 1 on relative velocity. Filtering distance values for range. MTI = 1 relative velocity exceeding Mth = 0 Mth = 0 • Filtering distance values for relative 20 km/h removes much noise. velocity exceeding ± 20 km/h removes much noise. Table A3. Radar parameter configurations. Table A3. Radar parameter configurations. Truck ID Common Parameters Specific Parameters Truck ID Common Parameters Specific Parameters Mode = 3 f = 5 MHz Mode = 3 BW = 200 MHz f0 = 5 MHz N = 200 samples BW = 200 MHz N = 5 targets tar Ns = 200 samples Ntar = 5 targets R = 100 m max f = 25 MHz MTI = 1 (active) 3 Rmax = 100 m N = 195 samples f0 = 25 MHz Mth = 0 (inactive) MTI = 1 (active) Ns = 195 samples Mth = 0 (inactive) Appendix A.2. Video Footage, Video Synchronization and Manual Video Coding Appendix A.2. Video Footage, Video Synchronization and Manual Video Coding Low-resolution video (LRV) file segments were converted to the MP4 format and merged using free Bandicut software [47]. Merged videos were imported to BORIS version Low-resolution video (LRV) file segments were converted to the MP4 format and 7.12.2. LRV files were used, since original MP4 files were too large for BORIS to handle. LRV merged using free Bandicut software [47]. Merged videos were imported to BORIS version files were 864 by 480 pixels, while original files were 1920 by 1080 pixels, both with 60 fps 7.12.2. LRV files were used, since original MP4 files were too large for BORIS to handle. frame rates. Conversion reduced the file size by an order of magnitude, while retaining LRV files were 864 by 480 pixels, while original files were 1920 by 1080 pixels, both with sufficient video quality for coding. Table A4 shows two code definitions, while examples of 60 fps frame rates. Conversion reduced the file size by an order of magnitude, while re- state events are shown in Table A5. taining sufficient video quality for coding. Table A4 shows two code definitions, while examples of state events are shown in Table A5. Table A4. Examples of video code definitions. Table A4. Examples of video code definitions. Video Code Definition Driving starts when the truck is fully inside the correct Video Code Definition lane on the roadway, with the steering wheel turned Driving starts when the truck is fully inside the correct lane on the roadway, with the steering wheel straight. It stops just before the driver starts turning the turned straight. It stops just before the driver starts turning the wheel, with the intention of entering Driving (S) wheel, with the intention of entering driveways, parking Driving (S) driveways, parking areas or stop pockets. Except during Break and periods of camera malfunctions, areas or stop pockets. Except during Break and periods of every other video code is coded only when Driving camera is also malfunctions, active. every other video code is coded only when Driving is also active. All time that is not Driving, is defined as Break. This includes maneuvering in and out of driveways, Break (S) All time that is not Driving, is defined as Break. This parking areas and stop pockets. Break (S) includes maneuvering in and out of driveways, parking areas and stop pockets. Table A5. Examples of coded state events for radar data filtering. Event Illustrations Tunnel Appl. Appl. Syst Syst . Innov. . Innov. 2023 2023 , 6 , 6 , x FO , x FO R P R P EER R EER R EVIE EVIE W W 23 of 23 of 29 29 dashcam and one Ns = 200 dashcam and one Ns = 200 • • Mode Mode 3 is p 3 is preferr referred ov ed ov erer mode 2; mode 2; • • Finalize Finalized parameters w d parameters were recom- ere recom- rad rad ar sensor ar sensor NN tartar = = 5 5 ma ma kes it kes it ea ea sy sy to to fifi ltler aw ter aw ay sta ay sta tit o ina ona ry ry mended to mended to avoid avoid inter interfefreence, y rence, yetet Rmax = 100 Rmax = 100 objects and objects and o o nn coming v coming v eh eh icles b icles b ase ase dd ma ma ximiz ximiz e sensor ra e sensor ra nge. nge. MTI MTI = 1 = 1 on relative velocity. on relative velocity. Mth = 0 Mth = 0 • • FiFi ltleri teri nn g di g di stst an an ce v ce v alu alu es for rela es for rela titi ve v e ve vlocit elocit yy exce exce ed ed ing ing ± ± 20 2 km/ 0 km/ hh remov remov es es much noise. much noise. Ta Ta bb le l A e A 3.3 Radar parameter con . Radar parameter con figurations figurations . . Truck ID Truck ID Common Common PP ar aa rmeters ameters Spec Spec ifi ifi c P c P arame arame te te rs r s Mode = 3 Mode = 3 f0f = 0 = 5 M 5 M H H z z 2 BW = 20 BW = 20 0 MH 0 MH z z NN s = 200 s = 200 sa sa m m pp lels es N N tar tar = 5 = 5 ta trget arget s s RR max max = 10 = 10 0 0 m m f0 = 25 MHz f0 = 25 MHz 3 3 MTI MTI = 1 ( = 1 ( act act iv iv e) e) NN s = 195 s = 195 sa sa m m pp lels es Mt Mt h h = 0 (in = 0 (in act act iv ie v) e ) Ap Ap pen pen did xi x A A .2..2 V . V ideo Foota ideo Foota ge, ge, V V ideo Sy ideo Sy nc nc hro hro nn izati izati on on an an d Manual d Manual V V ideo ideo C C odi odi ng ng Low-reso Low-reso lut lut ion video ion video (LR (LR VV ) ) fifi le le segme segme nn ts w ts w ere con ere con verted to the MP verted to the MP 4 form 4 form atat and and merged usin merged usin g free B g free B and and icut icut softw softw are are [47]. Merged v [47]. Merged v ide ide os w os w ere impo ere impo rted to BORIS v rted to BORIS v ersion ersion 7.7. 12 12 .2.2 . L . L RR VV fi fi les les w w ere ere use use dd , s , s ince o ince o rig rig in in alal MP MP 4 4fi fi lele s s ww ere t ere t oo oo lar lar ge for ge for BO BO RIS RIS t t o hand o hand le. le. LRV files were 864 by 480 pixels, while original files were 1920 by 1080 pixels, both with LRV files were 864 by 480 pixels, while original files were 1920 by 1080 pixels, both with 60 fps fr 60 fps fr ame rates. ame rates. Conv Conv er er sion sion red red uu ced ced the the fifi lele s s ize b ize b yy an o an o rder rder of of ma ma gnit gnit ud ud e,e, while while re re - - tatinin ainin g su g suffiffi cicent vid ient vid eo eo quality quality for co for co ding. ding. Ta Tble able AA 4 4shows two shows two c c ode de ode defifi nn itiitoin on s, wh s, wh ile ile example example s of state ev s of state ev ents ar ents ar e shown e shown in T in T able able AA 5. 5. Ta Ta bb le l A e A 4.4 Ex . Ex am am ples of ples of vi vi deo deo code code de de finition finition s. s. Vi Vi dd eo eo C C od od e eD D efi efi nition nition Driv Driv ing ing starts when th starts when th e e truck is fully truck is fully inside th inside th e correc e correc t lane on t lane on the the r r oadw oadw ay, with ay, with the the steerin steerin g g wheel wheel turne turne dd s s trta ria g iht. I ght. I t s t s top top s ju s ju st b st b efore efore the the driv driv er er stst ar ar tsts t t uu rning rning th th e w e w hh eel, eel, wi wi th th the the int int ent ent ion o ion o f e f e nn terin terin gg Driv Driv ing ing (S (S ) ) dri dri vv ewa ewa yy s,s, pa pa rki rki nn gg a r aea rea s s or st or st op pock op pock etet s.s Except . Except du du riri ng ng Br Br eak eak and periods of camera m and periods of camera m alfun alfun ction ction s, s, every o every o ther v ther v ideo cod ideo cod e is c e is c oded only wh oded only wh en en Driv Driv ing ing is i a s a lslo s a o a cti cti ve. ve. Appl. Syst. Innov. 2023, 6, 55 24 of 29 Al Al l ti l ti me t me t hh at a is t is not not Dri Dri ving ving , is defined , is defined as as Break Break . Th . Th is is in in clude clude s m s m ane ane uvering uvering in in an an d ou d ou t o t o f dr f dr ive ive w w ays ays , , Break Break (S (S ) ) pa pa rk rk in in gg a r aea rea s a s a nn d st d st op pock op pock etet s. s. Table A5. Examples of coded state events for radar data filtering. Ta Ta bb le l A e A 5.5 Exa . Exa m m ples of ples of cod cod ed e state d state ev ev ents fo ents fo r radar data r radar data fi lt fiering. ltering. Event Illustrations EvEvenent Il t Illuluststraratitions ons Tunnel Tunnel Tunnel Appl. Appl. Syst Syst . Innov. . Innov. 2023 2023 , 6 , 6 , x FO , x FO R P R P EER R EER R EE VIE VIE W W 24 of 24 of 29 29 Appl. Appl. Syst Syst . Innov. . Innov. 2023 2023 , 6 , 6 , x FO , x FO R P R P EER R EER R EE VIE VIE W W 24 of 24 of 29 29 Appl. Syst. Innov. 2023, 6, x FOR PEER REVIEW 24 of 29 Appl. Syst. Innov. 2023, 6, x FOR PEER REVIEW 24 of 29 Roundabout Roundabout Roundabout Roundabout Roundabout Roundabout Roundabout (Left): (Left (Left Straight; ): Straigh ): Straigh t; t; (Left (Left ): Straigh ): Straigh t; t; (Right) (Left Right-turn ): Straight; (Left): Straight; (R (R ight ight ) R ) R ight ight -t-u tu rn rn (R (R ight ight ) R ) R ight ight -t-u tu rn rn (Right) Right-turn (Right) Right-turn LWRR (likely LWRR (likely with with in in LWRR (likely LWRR (likely with with in in LWRR (likely within LWRR (likely within LWRR radar r radar r (likelyawithin nge) ange) radar range) radar r radar r ange) ange) radar r radar r ange) ange) LBR LBR RR (l (l ik iel kel y b y b eyon eyon dd LBR LBR RR (l (l ik iel kel y b y b eyon eyon dd LBRR (likely beyond LBRR (likely beyond radar r radar r ange) ange) LBRR (likely beyond radar range) radar r radar r ange) ange) radar range) radar range) Ap Ap pen pen dd ix ix A A .3 .3 . .R R adar Da adar Da ta C ta C uu rati rati on on Ap Ap pen pen dd ix ix A A .3 .3 . .R R adar Da adar Da ta C ta C uu rati rati on on Appendix A.3. Radar Data Curation Appendix A.3. Radar Data Curation Ra Ra dar dar d d aa ta ta p p oo st st -process -process in in g w g w aa s c s c aa rr rr ied ied out out in in Pyt Pyt hon hon 3. 3. 11 00 . The . The Pyt Pyt hh on libr on libr aa rires ies PP aa n- n- Appendix A.3. Radar Data Curation Ra Ra dar dar d d aa ta ta p p oo st st -process -process in in g w g w aa s c s c aa rr rr ied ied out out in in Pyt Pyt hon hon 3. 3. 11 00 . The . The Pyt Pyt hh on libr on libr aa rires ies PP aa n- n- Ra Ra dar dar d d aa ta ta p p oo st st -process -process in in g w g w aa s c s c aa rr rr ied ied out out in in Pyt Pyt hon hon 3. 3. 11 00 . The . The Pyt Pyt hh on libr on libr aa rires ies PP aa n- n- das, das, NumPy NumPy , datetime, Mat , datetime, Mat pp lotlib and lotlib and oo pp enpyxl enpyxl w w ee re re used used . All r . All r aa dd aa r d r d aa ta w ta w ee re re extracted extracted das, das, NumPy NumPy , datetime, Mat , datetime, Mat pp lotlib and lotlib and oo pp enpyxl enpyxl w w ee re re used used . All r . All r aa dd aa r d r d aa ta w ta w ee re re extracted extracted Radar data post-processing was carried out in Python 3.10. The Python libraries Pan- das, das, NumPy NumPy , datetime, Mat , datetime, Mat pp lotlib and lotlib and oo pp enpyxl enpyxl w w ere ere used used . All r . All r aa dd aa r d r d aa ta w ta w ere ere extracted extracted int intoo a P a P aand ndaas Dat s DataaFram Fram e.e. In mult In multi-t i-t ararget get scenar scenarios ios, object , objects w s w eerere place placedd suc suc cessiv cessiv eely ly das, NumPy int intoo a P a P , datetime, aand ndaas Dat s Dat Matplotlib aaFram Fram e.e. In mult In mult and openpyxl i-t i-t ararget get wer scenar scenar e used. ios ios,All object , object radar s w s w data eerere place wer place e extracted dd suc suc cessiv cessiv eely ly into a Pandas DataFrame. In multi-target scenarios, objects were placed successively into a Pandas DataFrame. In multi-target scenarios, objects were placed successively wi wi thi thi nn the Data the Data FF ra ra m m e, by de e, by de sc se ce nn dd ing SNR ing SNR . The Da . The Da ta ta Fr Fr aa m m e c e c oo ntai ntai ne ne dd the f the f oo lllo lo wing wing da da ta ta into a wi Pandas wi thi thi nn the Data the Data DataFrame. FF ra ra m In m e, by de e, by de multi-tar sc se c ge e nn d td iscenarios, ng SNR ing SNR . The Da . The Da objectsta wer ta Fr Fr a ea m placed m e c e c oo ntai ntai successively ne ne dd the f the f oo lwithin llo lo wing wing da da ta ta within the DataFrame, by descending SNR. The DataFrame contained the following data within the DataFrame, by descending SNR. The DataFrame contained the following data columns: Tim columns: Tim ee { { date date time time }, Dist }, Dist ance ance { fl {fl oa oa t}, t}, V V ee locit locit yy { {flfl oat oat },}, S S N N RR { {flfl oa oa t} t} and and Objec Objec t number t number the Dacolumns: Tim tcolumns: Tim aFrame, by descending ee { { date date time time }, Dist SNR. }, Dist The ance ance DataFrame { fl {fl oa oa t}, t}, V V ee locit contained locit yy { {flfl oat oat }, the }, S S N following N RR { {flfl oa oa t} t} and data and Objec Objec columns: t number t number columns: Time {datetime}, Distance {float}, Velocity {float}, SNR {float} and Object number columns: Time {datetime}, Distance {float}, Velocity {float}, SNR {float} and Object number {int {int }. Since r }. Since r aa dar timest dar timest am am ps did not cor ps did not cor respond respond t t oo lo lo cal t cal t ime ime , dat , dat ee titme shift ime shift s w s w ee re calc re calc u- u- Time {{datetime int {int }. Since r }. Since r }, Distance aa dar timest dar timest {float am }, am V ps did not cor ps did not cor elocity {float}, respond r SNR espond {float t t oo lo } lo and cal t cal t Ob ime ime ject , dat , dat number ee titme shift ime shift {int}. s w s w Since ee re calc re calc u- u- {int}. Since radar timestamps did not correspond to local time, datetime shifts were calcu- lat {int lat e}. Since r e dd ba ba sed sed o a o dar timest nn t t hh e prev e prev am iou iou ps did not cor sly sly correct correct ed ed rdate espond date an an d t d time time o lo of cal t of each ieach me , dat Rad Rad a er logg a tir logg me shift ing ing in in s w stst a ea n re calc n ce, ce, and and u- radar timestamps did not correspond to local time, datetime shifts were calculated based lat lat ee dd ba ba sed sed o o nn t t hh e prev e prev iou iou sly sly correct correct ed ed date date an an dd time time of of each each Rad Rad ar logg ar logg ing ing in in stst aa nn ce, ce, and and lat lat ed ed ba ba sed sed o o nn t t hh e prev e prev iou iou sly sly correct correct ed ed date date an an dd time time of of each each Rad Rad ar logg ar logg ing ing in in stst aa nn ce, ce, and and the the da da tete an an dd ti ti me me wi wi th whi th whi ch the ch the ra ra da da r output r output fifi les les w w ee re n re n aa med med (c (c f.f. T T aa ble ble A A 22 ).). Dat Dat ee titme ime on the previously corrected date and time of each Radar logging instance, and the date and the the da da tete an an dd ti ti me me wi wi th whi th whi ch the ch the ra ra da da r output r output fifi les les w w ee re n re n aa med med (c (c f.f. T T aa ble ble A A 22 ).). Dat Dat ee titme ime the date and time with which the radar output files were named (cf. Table A2). Datetime the date and time with which the radar output files were named (cf. Table A2). Datetime shifts w shifts w ee re re added to the T added to the T ime column, ime column, correct correct ing ing all all measurement measurement s. s. A A ll r ll r aa dd aa r d r d aa ta w ta w ee re re time with which the radar output files were named (cf. Table A2). Datetime shifts were shifts w shifts w ee re re added to the T added to the T ime column, ime column, correct correct ing ing all all measurement measurement s. s. A A ll r ll r aa dd aa r d r d aa ta w ta w ee re re shifts were added to the Time column, correcting all measurements. All radar data were shifts were added to the Time column, correcting all measurements. All radar data were merged merged into into one Dat one Dat aa Fr Fr ame and ame and sav sav ee dd as a P as a P ickle ickle fifi le le pr pr io io r to cu r to cu ra ra titi on. on. added to the Time column, correcting all measurements. All radar data were merged into merged merged into into one Dat one Dat aa Fr Fr ame and ame and sav sav ee dd as a P as a P ickle ickle fifi le le pr pr io io r to cu r to cu ra ra titi on. on. merged into one DataFrame and saved as a Pickle file prior to curation. merged into one DataFrame and saved as a Pickle file prior to curation. Filt Filt er er s s 4 (t 4 (t arg arg ee t sect t sect ion) ion) an an d 5 d 5 (down (down sa sa m m pling pling ) )are are ilil lust lust ra ra ted ted in F in F igur igur ee s s A2 A2 an an d d A A 33 . . one DataFrame and saved as a Pickle file prior to curation. Filt Filt er er s s 4 (t 4 (t arg arg ee t sect t sect ion) ion) an an d 5 d 5 (down (down sa sa m m pling pling ) )are are ilil lust lust ra ra ted ted in F in F igur igur ee s s A2 A2 an an d d A A 33 . . Filters 4 (target section) and 5 (downsampling) are illustrated in Figures A2 and A3. Fig Filt Fig u er u re s re 4 (t A2 A2 arg sh sh ows a e ows a t sect7-min ion) 7-min an period d 5 period (down for truck 2. Chosen for truck 2. Chosen sampling) are illust targets targets rated are b are b in F lue, d ilgur ue, d eis sc isc A2 arde arde an d d d ones ones A3. Filters 4 (target section) and 5 (downsampling) are illustrated in Figures A2 and A3. Fig Fig uu re re A2 A2 sh sh ows a ows a 7-min 7-min period period for truck 2. Chosen for truck 2. Chosen targets targets are b are b lue, d lue, d isc isc arde arde dd ones ones Fig Fig uu re re A2 A2 sh sh ows a ows a 7-min 7-min period period for truck 2. Chosen for truck 2. Chosen targets targets are b are b lue, d lue, d isc isc arde arde dd ones ones purple, purple, and t and t hh e moving e moving av av erage erage t t uu rq rq uo uo is ie. se. A A vv er er ag ag ing ing acro acro ss ss 1 1 00 log log gg ings ings , inst , inst ee aa dd of of few few ee r,r , purple, purple, and t and t hh e moving e moving av av erage erage t t uu rq rq uo uo is ie. se. A A vv er er ag ag ing ing acro acro ss ss 1 1 00 log log gg ings ings , inst , inst ee aa dd of of few few ee r,r , purple, purple, and t and t hh e moving e moving av av erage erage t t uu rq rq uo uo is ie. se. A A vv er er ag ag ing ing acro acro ss ss 1 1 00 log log gg ings ings , inst , inst ea ea dd of of few few er e,r , reduced reduced noise in noise in turbulen turbulen t situations. I t situations. I nn spec spec tion tion oo f rad f rad aa r dat r dat aa , e , e .g., .g., when when trav trav ersing ersing reduced reduced noise in noise in turbulen turbulen t situations. I t situations. I nn spec spec tion tion oo f rad f rad aa r dat r dat aa , e , e .g., .g., when when trav trav ersing ersing reduced noise in turbulent situations. Inspection of radar data, e.g., when traversing reduced noise in turbulent situations. Inspection of radar data, e.g., when traversing Roundabouts Roundabouts , , show show ed that ed that approxim approxim ately 10 ately 10 logg logg ing ing s w s w ee re nee re nee ded ded after the vid after the vid ee o cod o cod ee Roundabouts Roundabouts , , show show ed that ed that approxim approxim ately 10 ately 10 logg logg ing ing s w s w ee re nee re nee ded ded after the vid after the vid ee o cod o cod ee Roundabouts, showed that approximately 10 loggings were needed after the video code Roundabouts, showed that approximately 10 loggings were needed after the video code ended, ended, for for distance to the p distance to the p recedin recedin gg truc truc k to stab k to stab ilize ilize . . ended, ended, for for distance to the p distance to the p recedin recedin gg truc truc k to stab k to stab ilize ilize . . ended, for distance to the preceding truck to stabilize. ended, for distance to the preceding truck to stabilize. The presence The presence of of lone lone blue blue d d aa ta points loc ta points loc aa ted ted aw aw ay ay fro fro m m the b the b lue lue tre tre nn d in d in Fig Fig uu re re A2, A2, The presence The presence of of lone lone blue blue d d aa ta points loc ta points loc aa ted ted aw aw ay ay fro fro m m the b the b lue lue tre tre nn d in d in Fig Fig uu re re A2, A2, The presence of lone blue data points located away from the blue trend in Figure A2, shows tha shows tha The presence t thi t thi s sal al gori of gori lone thm thm blue ma ma y choose the wr d y choose the wr ata points locaong ted ong tar aw tar ay gg et. Er et. Er fromroneous se the b roneous se lue tre lection lection nd in sFig sare those are those ure A2, shows tha shows tha t thi t thi s sal al gori gori thm thm ma ma y choose the wr y choose the wr ong ong tar tar gg et. Er et. Er roneous se roneous se lection lection s sare those are those shows tha shows tha t thi t thi s sal al gori gori thm thm ma ma y choose the wr y choose the wr ong ong tar tar gg et. Er et. Er roneous se roneous se lection lection s sare those are those data points w data points w hh ich clear ich clear ly ly suppress the suppress the moving moving av av erage d erage d istance istance v v aa lue. Figure lue. Figure A3 A3 shows shows data points w data points w hh ich clear ich clear ly ly suppress the suppress the moving moving av av erage d erage d istance istance v v aa lue. Figure lue. Figure A3 A3 shows shows data points which clearly suppress the moving average distance value. Figure A3 shows data points which clearly suppress the moving average distance value. Figure A3 shows a a 10 10 -min -min exce exce rpt rpt of downs of downs aa mpled mpled ra ra da da r dat r dat aa (b (b lue lue ), ), alongside alongside dat dat aa after after fifi ltlt er er 1 1 ( ( gg ray ray ).) . a a 10 10 -min -min exce exce rpt rpt of downs of downs aa mpled mpled ra ra da da r dat r dat aa (b (b lue lue ), ), alongside alongside dat dat aa after after fifi ltlt er er 1 1 ( ( gg ray ray ).) . a 10-min excerpt of downsampled radar data (blue), alongside data after filter 1 (gray). a 10-min excerpt of downsampled radar data (blue), alongside data after filter 1 (gray). Appl. Syst. Innov. 2023, 6, 55 25 of 29 Figure A2 shows a 7-min period for truck 2. Chosen targets are blue, discarded ones purple, and the moving average turquoise. Averaging across 10 loggings, instead of fewer, reduced noise in turbulent situations. Inspection of radar data, e.g., when traversing Roundabouts, showed that approximately 10 loggings were needed after the video code ended, for distance to the preceding truck to stabilize. The presence of lone blue data points located away from the blue trend in Figure A2, shows that this algorithm may choose the wrong target. Erroneous selections are those Appl. Syst. Innov. 2023, 6, x FOR PEER REVIEW 25 of 29 data points which clearly suppress the moving average distance value. Figure A3 shows a Appl. Syst. Innov. 2023, 6, x FOR PEER REVIEW 25 of 29 10-min excerpt of downsampled radar data (blue), alongside data after filter 1 (gray). Figure A2. Truck 2 excerpt. Target selection when radar measures multiple objects simultaneously. Figure A2. Truck 2 excerpt. Target selection when radar measures multiple objects simultaneously. Figure A2. Truck 2 excerpt. Target selection when radar measures multiple objects simultaneously. Figure A3. Truck 2 excerpt following filters 1 and 5. Figure A3. Truck 2 excerpt following filters 1 and 5. Figure A3. Truck 2 excerpt following filters 1 and 5. Appendix A.4. Results and Discussion Appendix A.4. Results and Discussion Appendix Tabl A.4. es A6 Results –A13 show and Discussion statistics from each filtering step, the analysis of relative speed Tables A6–A13 show statistics from each filtering step, the analysis of relative speed and SNR as a function of distance bins, and radar operation in different driving situations. Tables A6–A13 show statistics from each filtering step, the analysis of relative speed and SNR as a function of distance bins, and radar operation in different driving situations. and SNR as a function of distance bins, and radar operation in different driving situations. Table A6. Relative velocity statistics (km/h) for trucks 2 and 3 after each filtering step. Table A6. Relative velocity statistics (km/h) for trucks 2 and 3 after each filtering step. Truck Number Truck Number Filtering Step 2 3 Filtering Step 2 3 Min Avg Max Std Min Avg Max Std Min Avg Max Std Min Avg Max Std Raw −280.9 −5.7 281.0 40.4 −280.7 −9.4 280.7 45.9 Raw −280.9 −5.7 281.0 40.4 −280.7 −9.4 280.7 45.9 1 −280.9 −7.7 281.0 43.7 −280.7 −11.3 280.7 48.9 1 −280.9 −7.7 281.0 43.7 −280.7 −11.3 280.7 48.9 2 −29.9 −0.5 29.9 5.5 −29.9 −0.5 29.9 5.5 2 −29.9 −0.5 29.9 5.5 −29.9 −0.5 29.9 5.5 3 −29.9 −0.3 29.9 4.8 −29.9 −0.3 29.7 4.6 3 −29.9 −0.3 29.9 4.8 −29.9 −0.3 29.7 4.6 4 −29.9 −0.2 29.9 4.4 −29.9 −0.2 29.7 4.2 4 −29.9 −0.2 29.9 4.4 −29.9 −0.2 29.7 4.2 5 −29.6 −0.2 28.6 4.3 −29.9 −0.2 27.2 4.0 5 −29.6 −0.2 28.6 4.3 −29.9 −0.2 27.2 4.0 6 −29.6 −0.1 28.6 4.2 −29.7 −0.1 27.2 3.8 6 −29.6 −0.1 28.6 4.2 −29.7 −0.1 27.2 3.8 Appl. Syst. Innov. 2023, 6, 55 26 of 29 Table A6. Relative velocity statistics (km/h) for trucks 2 and 3 after each filtering step. Truck Number Filtering Step 2 3 Min Avg Max Std Min Avg Max Std Raw 280.9 5.7 281.0 40.4 280.7 9.4 280.7 45.9 1 280.9 7.7 281.0 43.7 280.7 11.3 280.7 48.9 2 29.9 0.5 29.9 5.5 29.9 0.5 29.9 5.5 3 29.9 0.3 29.9 4.8 29.9 0.3 29.7 4.6 4 29.9 0.2 29.9 4.4 29.9 0.2 29.7 4.2 5 29.6 0.2 28.6 4.3 29.9 0.2 27.2 4.0 6 29.6 0.1 28.6 4.2 29.7 0.1 27.2 3.8 Table A7. SNR statistics (dB) for trucks 2 and 3 after each filtering step. Truck Number Filtering Step 2 3 Min Avg Max Std Min Avg Max Std Raw 6.8 24.6 53.9 10.8 6.6 22.6 51.4 9.7 1 6.9 27.6 53.9 10.6 6.7 24.2 51.4 9.9 2 6.9 29.9 53.9 9.9 6.8 28.1 51.4 8.5 3 15.1 31.8 53.9 8.3 15.1 29.5 51.4 7.3 4 15.1 32.9 53.9 7.9 15.1 30.4 51.4 7.0 5 15.1 31.6 49.7 6.9 15.1 29.4 47.7 5.9 6 15.1 31.8 49.7 6.9 15.1 29.5 47.7 5.8 Table A8. Average relative velocity (km/h) and SNR (dB) for truck 2 as a function of distance. Avg. Relative Velocity Distance Bins Average SNR (dB) # Measurements % of Total (km/h) 0–10 0.4 30.0 204 1% 10–20 1.0 32.3 1191 6% 20–30 0.3 31.1 3327 18% 30–40 0.1 31.8 5074 27% 40–50 0.0 32.4 5519 30% 50–60 0.2 31.4 2151 12% 60–70 0.1 31.2 813 4% 70+ 0.7 32.2 191 1% Table A9. Average relative velocity (km/h) and SNR (dB) for truck 3 as a function of distance. Avg. Relative Velocity Distance Bins Average SNR (dB) # Measurements % of Total (km/h) 0–10 0.3 29.4 199 1% 10–20 1.2 29.1 1113 6% 20–30 0.4 28.6 4173 23% 30–40 0.0 29.0 5642 31% 40–50 0.1 30.2 4988 28% 50–60 0.0 31.0 1502 8% 60–70 0.9 31.6 270 2% 70+ 0.4 31.4 29 0% Appl. Syst. Innov. 2023, 6, 55 27 of 29 Table A10. Relative velocity statistics (km/h) for trucks 2 and 3 during LWRR and Tunnel. Relative Velocity Truck Avg Min Max Std 2 0.1 19.7 28.6 3.5 3 0.1 19.3 18.8 3.3 Table A11. Distance (meters) and SNR (dB) statistics for trucks 2 and 3 during LWRR and Tunnel. Distance SNR Truck Avg Min Max Std Avg Min Max Std 2 36.4 10.6 68.2 10.5 30.8 15.1 46.9 6.4 3 36.2 13.8 65.0 9.6 28.5 15.7 44.0 5.2 Table A12. Distance (meters) and SNR (dB) statistics for trucks 2 and 3 during LWRR and Roundabout. Distance SNR Truck Avg Min Max Std Avg Min Max Std 2 23.3 8.6 58.4 10.1 24.5 16.2 39.5 5.2 3 22.4 6.3 60.0 11.2 22.4 15.1 41.5 4.4 Table A13. Relative velocity statistics (km/h) for trucks 2 and 3 during LWRR and Roundabout. Relative Velocity Truck Avg Min Max Std 2 6.5 27.8 11.2 9.1 3 5.9 25.7 10.5 8.9 References 1. Tsugawa, S.; Jeschke, S.; Shladover, S.E. A Review of Truck Platooning Projects for Energy Savings. IEEE Trans. Intell. Veh. 2016, 1, 68–77. [CrossRef] 2. Eitrheim, M.H.R.; Log, M.M.; Tørset, T.; Levin, T.; Pitera, K. Opportunities and Barriers for Truck Platooning on Norwegian Rural Freight Routes. Transp. Res. Rec. 2022, 2676, 810–824. [CrossRef] 3. Horenberg, D. Applications within Logistics 4.0: A Research Conducted on the Visions of 3PL Service Providers. In Proceedings of the 9th IBA Bachelor Thesis Conference; University of Twente, The Faculty of Behavioural, Management and Social Sciences: Enschede, The Netherlands, 2017. 4. Bergenhem, C.; Shladover, S.; Coelingh, E.; Englund, C.; Tsugawa, S. Overview of Platooning Systems. In Proceedings of the 19th ITS World Congress, Vienna, Austria, 22–26 October 2012. 5. Konstantinopoulou, L.; Coda, A.; Schmidt, F. Specifications for Multi-Brand Truck Platooning. In Proceedings of the 8th International Conference on Weigh-In-Motion, Prague, Czech Republic, 20–23 May 2019. 6. Borealis. Available online: https://www.vegvesen.no/vegprosjekter/europaveg/e8borealis/ (accessed on 1 March 2023). 7. Simonsen, J. Ground-Breaking EU Project on Automated Heavy-Haul Freight Vehicles to Be Launched. ITS Norway: Oslo, Norway. Press Release, 1 October 2022. Available online: https://its-norway.no/ground-breaking-eu-project-on-automated- heavy-haul-freight-vehicles-to-be-launched/ (accessed on 3 January 2023). 8. Robinson, T.; Coelingh, E. Operating Platoons On Public Motorways: An Introduction To The SARTRE Platooning Programme. In Proceedings of the 17th World Congress on Intelligent Transport Systems, Busan, Republic of Korea, 25–29 October 2010. 9. Zhang, L.; Chen, F.; Ma, X.; Pan, X. Fuel Economy in Truck Platooning: A Literature Overview and Directions for Future Research. J. Adv. Transp. 2020, 2020, e2604012. [CrossRef] 10. Srisomboon, I.; Lee, S. Efficient Position Change Algorithms for Prolonging Driving Range of a Truck Platoon. Appl. Sci. 2021, 11, 10516. [CrossRef] 11. Hakobyan, G.; Yang, B. High-Performance Automotive Radar: A Review of Signal Processing Algorithms and Modulation Schemes. IEEE Signal Process. Mag. 2019, 36, 32–44. [CrossRef] 12. Scheiner, N.; Kraus, F.; Appenrodt, N.; Dickmann, J.; Sick, B. Object Detection for Automotive Radar Point Clouds—A Comparison. AI Perspect. 2021, 3, 6. [CrossRef] Appl. Syst. Innov. 2023, 6, 55 28 of 29 13. Ju, Y.; Jin, Y.; Lee, J. Design and Implementation of a 24 GHz FMCW Radar System for Automotive Applications. In Proceedings of the 2014 International Radar Conference, Lille, France, 13–17 October 2014; pp. 1–4. 14. Venon, A.; Dupuis, Y.; Vasseur, P.; Merriaux, P. Millimeter Wave FMCW RADARs for Perception, Recognition and Localization in Automotive Applications: A Survey. IEEE Trans. Intell. Veh. 2022, 7, 533–555. [CrossRef] 15. Bilik, I.; Longman, O.; Villeval, S.; Tabrikian, J. The Rise of Radar for Autonomous Vehicles: Signal Processing Solutions and Future Research Directions. IEEE Signal Process. Mag. 2019, 36, 20–31. [CrossRef] 16. Shnidman, D.A. Radar Detection in Clutter. IEEE Trans. Aerosp. Electron. Syst. 2005, 41, 1056–1067. [CrossRef] 17. Wang, L.; Giebenhain, S.; Anklam, C.; Goldluecke, B. Radar Ghost Target Detection via Multimodal Transformers. IEEE Robot. Autom. Lett. 2021, 6, 7758–7765. [CrossRef] 18. Ortiz, F.M.; Sammarco, M.; Costa, L.H.M.K.; Detyniecki, M. Applications and Services Using Vehicular Exteroceptive Sensors: A Survey. IEEE Trans. Intell. Veh. 2023, 8, 949–969. [CrossRef] 19. Gu, Y.; Hsu, L.-T.; Kamijo, S. Passive Sensor Integration for Vehicle Self-Localization in Urban Traffic Environment. Sensors 2015, 15, 30199–30220. [CrossRef] [PubMed] 20. Wang, F.; Zhuang, W.; Yin, G.; Liu, S.; Liu, Y.; Dong, H. Robust Inter-Vehicle Distance Measurement Using Cooperative Vehicle Localization. Sensors 2021, 21, 2048. [CrossRef] [PubMed] 21. Kim, T.-W.; Jang, W.-S.; Jang, J.; Kim, J.-C. Camera and Radar-Based Perception System for Truck Platooning. In Proceedings of the 2020 20th International Conference on Control, Automation and Systems (ICCAS), Busan, Republic of Korea, 13–16 October 2020; pp. 950–955. 22. URAD Radar for Raspberry Pi. Available online: https://urad.es/en/product/urad-radar-raspberry-pi/ (accessed on 26 January 2023). 23. Eitrheim, M.H.R.; Log, M.M.; Tørset, T.; Levin, T.; Nordfjærn, T. Driver Workload in Truck Platooning: Insights from an on-Road Pilot Study on Rural Roads. In Proceedings of the 32nd European Safety and Reliability Conference (ESREL 2022), Dublin, Ireland, 28 August–1 September 2022. 24. Knoop, V.L.; Wang, M.; Wilmink, I.; Hoedemaeker, D.M.; Maaskant, M.; Van der Meer, E.-J. Platoon of SAE Level-2 Automated Vehicles on Public Roads: Setup, Traffic Interactions, and Stability. Transp. Res. Rec. 2019, 2673, 311–322. [CrossRef] 25. Mills, D.L. Internet Time Synchronization: The Network Time Protocol. IEEE Trans. Commun. 1991, 39, 1482–1493. [CrossRef] 26. Emerald Sequoia LLC Emerald Time. Available online: https://emeraldsequoia.com/et/ (accessed on 27 January 2023). 27. Racelogic 04—VBOX Sport Logging. Available online: https://en.racelogic.support/VBOX_Motorsport/Product_Info/ Performance_Meters/VBOX__Sport/VBOX_Sport_User_Guide/04_-_VBOX_Sport_Logging (accessed on 8 March 2023). 28. Rohling, H.; Meinecke, M.-M. Waveform Design Principles for Automotive Radar Systems. In Proceedings of the 2001 CIE International Conference on Radar Proceedings (Cat No.01TH8559), Beijing, China, 15–18 October 2001; pp. 1–4. 29. Al-Hasan, T.M.; Shibeika, A.S.; Attique, U.; Bensaali, F.; Himeur, Y. Smart Speed Camera Based on Automatic Number Plate Recognition for Residential Compounds and Institutions Inside Qatar. In Proceedings of the 2022 5th International Conference on Signal Processing and Information Security (ICSPIS), Dubai, United Arab Emirates, 7–8 December 2022; pp. 42–45. 30. Murad, M.; Bilik, I.; Friesen, M.; Nickolaou, J.; Salinger, J.; Geary, K.; Colburn, J.S. Requirements for next Generation Automotive Radars. In Proceedings of the 2013 IEEE Radar Conference (RadarCon13), Ottawa, ON, Canada, 29 April–3 May 2013; pp. 1–6. 31. Ouaknine, A.; Newson, A.; Rebut, J.; Tupin, F.; Pérez, P. CARRADA Dataset: Camera and Automotive Radar with Range- Angle- Doppler Annotations. In Proceedings of the 2020 25th International Conference on Pattern Recognition (ICPR), Milan, Italy, 10–15 January 2021; pp. 5068–5075. 32. Radhityo Sardjono, D.; Suratman, F.Y. Istiqomah Human Motion Change Detection Based On FMCW Radar. In Proceedings of the 2022 IEEE Asia Pacific Conference on Wireless and Mobile (APWiMob), Bandung, Indonesia, 9–10 December 2022; pp. 1–6. 33. Erdyarahman, R.; Suratman, F.Y.; Pramudita, A.A. Contactless Human Respiratory Frequency Monitoring System Based on FMCW Radar. In Proceedings of the 2022 IEEE Asia Pacific Conference on Wireless and Mobile (APWiMob), Bandung, Indonesia, 9–10 December 2022; pp. 1–7. 34. Oberhammer, J.; Somjit, N.; Shah, U.; Baghchehsaraei, Z. 16—RF MEMS for Automotive Radar. In Handbook of Mems for Wireless and Mobile Applications; Uttamchandani, D., Ed.; Woodhead Publishing Series in Electronic and Optical Materials; Woodhead Publishing: Sawston, UK, 2013; pp. 518–549; ISBN 978-0-85709-271-7. 35. Friard, O.; Gamba, M. BORIS: A Free, Versatile Open-Source Event-Logging Software for Video/Audio Coding and Live Observations. Methods Ecol. Evol. 2016, 7, 1325–1330. [CrossRef] 36. Norwegian Public Roads Administration (NPRA) Handbook N302 Road Markings. Available online: https://viewers.vegnorm. vegvesen.no/product/859926/nb#id-8d46a74b-9c46-4790-b012-974ebf080aed (accessed on 20 February 2023). 37. Suzuki, H.; Nakatsuji, T. Dynamic Estimation of Headway Distance in Vehicle Platoon System under Unexpected Car-Following Situations. Transp. Res. Procedia 2015, 6, 172–188. [CrossRef] 38. Kim, T.; Park, T.-H. Extended Kalman Filter (EKF) Design for Vehicle Position Tracking Using Reliability Function of Radar and Lidar. Sensors 2020, 20, 4126. [CrossRef] [PubMed] 39. Jing, L.; Yanping, Z.; Xingang, Z. On the Maximum Unambiguous Range of LFMCW Radar. In Proceedings of the 2019 International Conference on Information Technology and Computer Application (ITCA), Guangzhou, China, 20–22 December 2019; pp. 92–94. Appl. Syst. Innov. 2023, 6, 55 29 of 29 40. Berthold, P.; Michaelis, M.; Luettel, T.; Meissner, D.; Wuensche, H.-J. Radar Reflection Characteristics of Vehicles for Contour and Feature Estimation. In Proceedings of the 2017 Sensor Data Fusion: Trends, Solutions, Applications (SDF), Bonn, Germany, 10–12 October 2017; pp. 1–6. 41. Macaveiu, A.; Câmpeanu, A. Automotive Radar Target Tracking by Kalman Filtering. In Proceedings of the 2013 11th International TEL Conference on Telecommunications in Modern Satellite, Cable and Broadcasting Services ( SIKS), Nis, Serbia, 16–19 October 2013; Volume 2, pp. 553–556. 42. Folster, F.; Rohling, H. Data Association and Tracking for Automotive Radar Networks. IEEE Trans. Intell. Transp. Syst. 2005, 6, 370–377. [CrossRef] 43. Scheiner, N.; Appenrodt, N.; Dickmann, J.; Sick, B. A Multi-Stage Clustering Framework for Automotive Radar Data. In Proceedings of the 2019 IEEE Intelligent Transportation Systems Conference (ITSC), Auckland, New Zealand, 27–30 October 2019; pp. 2060–2067. 44. Schlichenmaier, J.; Roos, F.; Kunert, M.; Waldschmidt, C. Adaptive Clustering for Contour Estimation of Vehicles for High- Resolution Radar. In Proceedings of the 2016 IEEE MTT-S International Conference on Microwaves for Intelligent Mobility (ICMIM), San Diego, CA, USA, 19–20 May 2016; pp. 1–4. 45. Domhof, J.; Kooij, J.F.P.; Gavrila, D.M. A Joint Extrinsic Calibration Tool for Radar, Camera and Lidar. IEEE Trans. Intell. Veh. 2021, 6, 571–582. [CrossRef] 46. Ji, Z.; Prokhorov, D. Radar-Vision Fusion for Object Classification. In Proceedings of the 2008 11th International Conference on Information Fusion, Cologne, Germany, 30 June–3 July 2008; pp. 1–7. 47. Bandicam Company Bandicut Video Cutter, Joiner and Splitter Software. Available online: https://www.bandicam.com/ bandicut-video-cutter/ (accessed on 27 January 2023). Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Journal

Applied System Innovation – Multidisciplinary Digital Publishing Institute

Published: May 6, 2023

Keywords: inter-vehicle distance measurements; radar sensor; action camera; field study; automated truck platooning; rural road; manual video coding; field-of-view; tunnels; roundabouts

There are no references for this article.