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

Learn More →

Using a Fuzzy-Bayesian Approach for Predicting the QoS in VANET

Using a Fuzzy-Bayesian Approach for Predicting the QoS in VANET Applied Computer Systems ISSN 2255-8691 (online) ISSN 2255-8683 (print) December 2022, vol. 27, no. 2, pp. 101–109 https://doi.org/10.2478/acss-2022-0011 https://content.sciendo.com Using a Fuzzy-Bayesian Approach for Predicting the QoS in VANET 1* 2 3 4 Hafida Khalfaoui , Abdellah Azmani , Abderrazak Farchane , Said Safi 1,3,4 LIMATI Laboratory, Sultan Moulay Slimane University, Beni Mellal, Morocco Intelligent Automation Laboratory, Abdelmalek Essaadi University, Tetouan, Morocco Abstract – There are considerable obstacles in the transport implement a tool for predicting problems that disrupt the proper sector of developing countries, including poor road conditions, functioning of the network. Therefore, this approach allows poor road maintenance and congestion. The dire impacts of these studying the influence of input parameters, including routing, challenges could be extremely damaging to both human lives and security and data processing, on the QoS in VANET. After that, the economies of the countries involved. Intelligent Transportation the risk can be resolved or even eliminated to safeguard system Systems (ITSs) integrate modern technologies into existing connectivity and avoid the problems of disruption in the transportation systems to monitor traffic. Adopting Vehicular Ad- hoc Network (VANET) into the road transport system is one of the transport, generating additional costs, distribution delays and most ITS developments demonstrating its benefits in reducing ultimately dissatisfaction. incidents, traffic congestion, fuel consumption, waiting times and The remainder of this article is structured as follows: pollution. However, this type of network is vulnerable to many Section II shows the challenges of VANET. Section III presents problems that can affect the availability of services. This article the literature related to the QoS in VANET. Section IV explains uses a Fuzzy Bayesian approach that combines Bayesian Networks the construction of the FB model. Section V discusses the (BN) and Fuzzy Logic (FL) for predicting the risks affecting the quality of service in VANET. The implementation of this model approach proposed. The final section concludes this work and can be used for different types of predictions in the networking presents some future areas of research. field and other research areas. II. CHALLENGES OF VANET Keywords – Bayesian network, fuzzy-Bayesian, fuzzy logic, prediction, quality of service, risk analysis, VANET. Significant problems appear when using ad-hoc communication between vehicles moving in a road I. INTRODUCTION environment. The key challenges are as follow: VANET is a Mobile Ad-hoc Network (MANET) subclass. A. Routing Problems Its objective is to create a communication system for Vehicle- The routing protocol that establishes and maintains routes to-Vehicle (V2V), Vehicle-to-Infrastructure (V2I) and from source to destination is the most challenging issue in Infrastructure to-Infrastructure (I2I) that increases awareness in VANET [5]. Since the infrastructure of VANET is limited, the the route and allows acquiring real-time traffic events, such as routing mechanism must be executed by the vehicles accidents, pre-crash warnings and emergency electronic brake themselves, and the intermediate route packets sent between lights [1], [2]. The mobile nodes are smart vehicles supplied distant vehicles [6]. This mechanism suffers from many issues with flexible computing resources containing computers, due to the unique features of VANET, such as the dynamic geolocation systems (GPS), radars, network devices and topology caused by high mobility of vehicles and variable various sensors [3]. These vehicles are interconnected and can network density. Furthermore, the cars may join and quit the have an interface that allows the driver to receive and send network rapidly, causing frequent path disruption. messages to various network entities. Furthermore, they Consequently, developing an efficient protocol that keeps a incorporate a positioning device such as GPS and sensors that route is difficult [7]. Besides, some nodes act selfishly and detect changed conditions, such as the state of the trailer’s doors refuse to participate in the routing to save their resources, such (open or closed), the temperature inside, violent movements as storage and processing time [6], which can degrade and harm and non-programmed stops [4]. These characteristics give high the performance of the network [7]. There are additional issues visibility and control of different operations and activities that regarding the VANET routing protocols. The most common occur within the transportation and help predict risks and make are: the scalability and limited bandwidth leading to congestion right decisions. and interference in the network [8], signal transmission This paper aims at measuring the quality of service (QoS) in attenuations caused by adverse weather and obstacles, also VANET by implementing a Fuzzy-Bayesian (FB) approach. security issues caused by malicious nodes that can launch The latter starts by analysing the risks of VANET in order to Corresponding author’s e-mail: hafidakhalfaoui1996@gmail.com ©2022 Hafida Khalfaoui, Abdellah Azmani, Abderrazak Farchane, Said Safi. This is an open access article licensed under the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0). 101 Applied Computer Systems _________________________________________________________________________________________________2022/27 diverse attacks, such as a black hole, eavesdropping, wormhole Availability and rushing attacks [9]. Availability is a critical factor that aims at protecting the life of users in VANET. It ensures that the network is operational B. Security Problems and useful information is always available [18]. The most The security aspect is an important pillar in VANET. If an famous availability threats are denial of service (DOS) attacks, attacker arrives to launch an attack, it can have a significant jamming attacks, malware attacks, greedy behaviour attacks, impact on network operations and services. Passive attacks, spamming attacks and so on. Their common goals are to use the such as listening to the communication and using the secret data bandwidth unnecessarily to cause voluntary collisions, decrease to execute malicious activities, or active attacks, such as communication quality between vehicles and thus interrupt removing, modifying, or repeating messages to disrupt the services or prohibit other nodes from using VANET support network, are both possible. It can also program traffic signals to and services [17]. cause traffic jams or block streets [10]. VANET presents several security issues because of its unique characteristics, C. Data Processing Problems such as the open network environment and dynamic network In recent years, the VANET environment has been topology [11]. The following are the most notable security generating massive amounts of data due to sensor and threats. communication technologies. These data are very important in improving productivity, travel comfort, transportation safety Authentication and modern society’s economic prosperity. As the amount of In VANET, there are two types of authentication: identity data generated by VANET grows exponentially, its complexity authentication, which determines the legitimacy of nodes, and and processing will present immense challenges due to the lack message authentication, which prevents fraudulent of intelligent systems that manage the large quantities of data transmission of messages in the network. Several existing generated. authentication algorithms have solved some security problems As a result, VANET requires complementary technologies to in VANETs, but their security still needs improvement. help optimize and efficiently process all data collected from Impersonation, identity (ID) disclosure attacks, spoofing disparate sources. Big data and artificial intelligence (AI) attacks, etc., are types of authentication attacks. They are provide competitive advantages in this context, allowing generated when malicious nodes can create false identities or organisations to make better decisions, while reducing reveal other node identities to act as legitimate vehicles and operational time and costs. However, these technologies present benefit from or reinforce incorrect data in the network. For significant challenges in terms of energy, node storage capacity example, an attacker may pretend to be an emergency vehicle and computation costs [19]. With the recent emergence of new in order to slow down other cars [12]. integrated architectures and access technologies, the Privacy Protection computation burden of IA algorithms and big data can be alleviated by migrating complex computations to external cloud Privacy protection has extended concern for VANET. servers [20]. This solution also presents some problems in terms Vehicles in contact with it typically share messages without any of time and security. When the scalability of nodes increases, privacy or security, which means the message is not encrypted the existing cloud computing can hardly satisfy the needs of end and contains vehicle ID, speed, location and other specific users. In general, uploading to and downloading from the cloud information that is usually linked to the driver’s identity. If consume time and energy [21]. these private data are open to attackers, this latter can use them to profile users or launch various attacks, such as masquerading III. RELATED WORK and impersonation attacks that can mislead other vehicles and endanger road safety [13], [14]. QoS measures the service degree of satisfaction as presented to the end user. In telecommunication networks, QoS aims at Confidentiality achieving better communication behaviour by routing the Confidentiality is a necessary security requirement for packets correctly and using resources optimally [22]. exchanging safety and non-safety messages between vehicles. Providing QoS support to VANET is an active research area. The message content should be secure and inaccessible to non- This network has particular features, such as high mobility, authenticated users. Confidentiality also prevents malicious scalability, a lack of centralized control unit, a hidden node activities, such as man-in-the-middle and traffic analysis issue, limited resources and an insecure medium, which make attacks [15], [16]. QoS provisioning very complex. In VANET, each application has specific QoS requirements; Integrity for example, a safety application should have a minimum End- Data integrity is one of the VANET security services that to-End Delay (E2ED) and packet loss because if a warning focus on conserving and ensuring the precision and accuracy of message does not arrive at a destination quickly, it will not help data during their transmission. There are numerous ways in prevent an accident [23]. which the data integrity service can be compromised. The most In the literature, many efforts and studies are dedicated to the common data integrity threats are replay, message tampering factors that affect the QoS in VANET and the majority are and illusion attacks [17]. looking into the routing aspect. Applied Computer Systems _________________________________________________________________________________________________2022/27 The authors of [24] proposed QMM-VANET to maintain E2ED of data transmission and, consequently, on the routing security and connectivity in the VANET. It is a clustering quality in the network? routing protocol that considers QoS requirements and mobility Fig. 1 represents a hierarchical arrangement, with both constraints. Firstly, this protocol calculates the QoS value of mobility and density are being ancestors and parents of E2ED; each vehicle and chooses a trustworthy one as the cluster head. analogical routing is its descendant and its child. Then, it selects gateways among neighbouring nodes and uses This study started with exhaustive research to identify the a gateway recovery algorithm to pick another gateway if link main challenges of a VANET network as follows: fails. In comparison to other protocols, this one demonstrated • Risks related to data routing influenced by the route its efficacy in terms of packet delivery ratio and E2ED in the conditions, such as mobility, density of nodes and highway scenario. security attacks. In [25] and [26], the authors conduct brief analyses on QoS, • Risks related to the security of the data circulated in the which include some types of routing protocols that help network due to the lack of robust mechanisms, such as minimize delay and increase overall QoS in VANET. The authentication, confidentiality, privacy and integrity. authors of [27] also improved various QoS parameters for • Risks related to data processing facilitating to make various scenario networks, such as delay, jitter, throughput and decisions in the different activities. Processing is affected packet loss ratio, by employing various routing algorithms with by the quality and quantity of data collected from the different mobility patterns, adaptive modulation, multiple-input different entities of the network. multiple-output (MIMO) and coding (AMC) techniques. This study also identified all the parameters that could In [28], the authors exploit Road Side Units (RSUs) to influence the occurrence degree of the above-mentioned risks. accumulate and transmit traffic data. They attempt to find the This identification led to conceiving the nature of the shortest route to the destination based on the cost of each dependencies between the different events and defining the transmission path using an effective algorithm called Dijkstra’s causal architecture of the Bayesian model. The following steps Dedicated Short Range Communication (DDSRC). are proposed for the analysis of QoS in VANET developed by According to the authors of [29], if the number of packets the BN: transmitted in the network increases, a queue of packets may be Step 1: Define the architecture of the BN (causal graph). created in the vehicle, and the response time of some packets Step 2: Generate conditional probabilities of intermediate may expire before being processed, causing a decreased service effects and final impacts. rate. They attempted to improve VANET performance by A. Architecture of the Bayesian Network of QoS in VANET removing useless or unused packets, in contrast to related works The architecture of the BN is based on the identification of that investigated the increasing service rate by changing the the relations between the nodes constituting the graph, which is parameters and properties of scheduler algorithms. divided according to their typology into three classes: In overall, it can be concluded that none of the previous QoS • The input parameters of the graph: these parameters solutions met all of the QoS criteria. represent the input nodes of the network. • Intermediate effects: the direct effects that are broken IV. MODELLING THE RISK OF QOS IN VANET USING A FUZZY- down into input parameters that are the origin of the BAYESIAN APPROACH causes. A Bayesian Network (BN) is a probabilistic graphical model • Final impacts: the factors that directly influence the QoS developed by Pearl to aid in reasoning under uncertainty [30]. in VANET. BN is defined by a pair (G, O) with G = V, A is a directed The investigation of the various factors that influence QoS in acyclic graph that encodes a joint probability distribution over VANETs aids in identifying input parameters, as described in a finite set of categorical variables V, and the arcs A represent Table I, and intermediate effects and final impacts, as shown in direct relationships between them [31]. In addition, a set of Table II. The causal graph in Fig. 2 is developed based on the parameters O defines the behaviour of each variable caused by two tables and after studying the nature of the causal its parent in the graph. dependencies between them. In this study, this tool allows for the prediction of problems that disturb the proper functioning of the network by studying B. Generation of Conditional Probabilities of Intermediate Effects and Final Impacts the influence of input parameters on such problems. For example, what is the impact of mobility and density on the After constructing the BN graph, the conditional probabilities (CPs) for each variable must be computed. There are numerous sources of probabilistic information available, such as databases containing road traffic and attacks. In this study, the BN considers a large number of parameters that make the existence of a complete database a difficult task. Unfortunately, the available databases are private and Fig. 1. Example of a causal relationship. insufficiently rich to permit a reliable examination of the https://www.unb.ca/cic/datasets/index.html Applied Computer Systems _________________________________________________________________________________________________2022/27 needed probabilities [47]. Therefore, FL is proposed for lies in its resemblance to human reasoning and natural generating these CPs systematically. It is an effective solution language. for handling imprecise data and linguistic problems. Its strength TABLE I DESCRIPTION OF THE INPUT PARAMETERS OF THE CAUSAL GRAPH CATEGORY OF VARIABLE NAME DESCRIPTION VARIABLES Security Authentication It represents the mechanism used to provide authenticity of both the data origin and also verify its sender. Integrity It represents the mechanism used to ensure that a sent message should reach the destination on the chosen path without any alteration [32]. Confidentiality The mechanism ensures that only authorized users can access the data. Failure to meet this requirement compromises the security of the exchanged data and the privacy of users [33]. Privacy This mechanism guarantees the anonymity of drivers and prevents an unauthorized party from knowing the combination between the sender’s real identity and private information related to traffic messages, such as vehicle identity, direction and location [34]. Inadvertence It refers to threats that result from trusted insiders who do not have malicious intent but fail to manage security responsibly. As a result, a malicious outsider is frequently able to use the privileges of the inattentive insider to carry out security problems. Selfish behaviour It refers to selfish nodes in the network that may refuse to cooperate in forwarding messages in order to increase their own resources [35]. Routing Mobility It refers to the movement of nodes communicating wirelessly and varies in function of their speed. Density It is the number of vehicles driving concurrently on the road [36]. The routing protocol is used to maintain links in the network to perform packet forwarding. In the VANET, Type of protocol there are many types of routing protocols. Choosing an effective one is very difficult because it depends on various parameters, such as mobility model, environment and others [37]. Bandwidth It is the maximum throughput at which packets can be transmitted between nodes without disrupting any existing network flow [38]. Frequency It is the number of occurrences of a repeating event per unit of time. If the frequency is high, we get short-range communication, which results in poor connectivity [39]. Obstacle Things that can diffract or block the signals in the road, such as buildings, trees, hills and bridges [40]. Weather Adverse weather conditions provide a signal transmission attenuation, like rainy or snowy days and dusty weather. Therefore, they will negatively influence the connectivity of high-speed data links in the vehicle communication range [41]. Data processing Internet Amount of data uploaded from the internet that can improve the productivity, travel comfort and modern society’s economic prosperity. OBU Amount of data exchanged with OBU. It concludes V2V and V2I communications. Sensors Amount of data collected from sensor systems related to state of vehicle or the external environment (based on radar systems, video cameras, image processing, etc.) [42]. Type of It is critical to choose the suitable type of data processing to avoid negative effect on the data output or the final processing product. Device resources These are resources/components that allow vehicle to act as a mobile service provider of sensing, data storage, data relaying, computing, infotainment and localization services [43]. TABLE II DESCRIPTION OF INTERMEDIATE EFFECTS AND FINAL IMPACTS INTERMEDIATE DESCRIPTION EFFECTS Signal transmission It represents the quality of signal used to transmit data between nodes. Interference Interference is caused by any signal other than the one configured in the network, and it obstructs the network normal operation. It usually causes slower speeds, higher latency, frequent disconnects, and sometimes, a complete failure to connect. Congestion This variable describes the state of the network when there is an overload of the links, which progressively affects the network performance with an increase in the transmission delay and a decrease in throughput [44]. Connectivity This property ensures successful data transmission from a source vehicle to a destination. It can be obtained if the number of active vehicles is increased in the network [45]. E2ED The time required to send a packet from its origin to its destination [46]. Packet loss The number of packets that fail to reach their destination. Applied Computer Systems _________________________________________________________________________________________________2022/27 Attacks It represents the number of attacks launched in the network. Security It represents the level of security in the network. Data processing Data processing is the process of collecting data and converting it into usable information (graphs, documents, etc.), to be interpreted by computers using various AI algorithms. This variable represents the quality of data processing used in the VANET. Quantity of data The amount of data generated in the network due to sensor and communication technologies. FINAL IMPACTS DESCRIPTION Routing This variable represents the quality of routing that has a great effect in the transmission of data between nodes. Quality of data Data quality is determined by many factors such as accuracy, completeness, confidentiality, validity, uniqueness and timeliness. QoS in VANET QoS measures the service degree of satisfaction as presented to the end user in the VANET. Fig. 2. BN modelling the analyses of risk of QoS in VANET. The implementation of the FL for generating CPs is done in TABLE III three principal steps [48], [49]: STATES OF THE BN NODES • Fuzzification: converts the input variables into a fuzzy VARIABLE LINGUISTIC VALUES subset using fuzzy linguistic values and membership functions. Authentication Robust, medium, weak • Inference: evaluates and combines the fuzzy rules to form Integrity Robust, medium, weak conclusions that give the fuzzy outputs of the system. Confidentiality Robust, medium, weak • Defuzzification: transforms the conclusions provided by Privacy Robust, medium, weak the inference engine into numeric values representing the Inadvertence High, medium, low final response of the fuzzy system. The most widely used inference methods are these of Attacks High, medium, low Mamdani [50] and Sugeno [39]. The main difference between Selfish behaviour High, medium, low them lies in the way the crisp output is obtained from the fuzzy Security Good, medium, bad inputs. Mamdani uses a defuzzification technique of fuzzy Frequency High, medium, low outputs, while Sugeno uses a weighted average to calculate the result values [51]. This study chooses the method of Sugeno Bandwidth Large, medium, small because it has a better processing time. As a result, generating Congestion High, medium, low CPs is done just with two first steps. Mobility High, medium, low Fuzzification Density High, medium, low To implement the fuzzification step, the fuzzy variables and Interferences High, medium, low their linguistic values are presented in Table III. Obstacle Many, medium, few Weather Good, medium, bad Applied Computer Systems _________________________________________________________________________________________________2022/27 TABLE IV Type of protocol Good, medium, bad STATES OF THE BN NODES Signal transmission Good, medium, bad RULE IF QUALITY OF AND THEN QOS IN Connectivity Good, medium, bad DATA ROUTING VANET E2ED Long, medium, short 1 medium bad bad Packet loss High, medium, low 2 medium medium medium OBU Large, medium, small 3 medium good medium Sensors Large, medium, small 4 bad bad bad Internet Large, medium, small 5 bad medium bad Quantity of data Large, medium, small 6 bad good bad Type of processing Good, medium, bad 7 good bad bad Devices resources High, medium, limited 8 good medium medium Data processing Good, medium, bad 9 good good good Quality of data Good, medium, bad Routing Good, medium, bad QoS in VANET Good, medium, bad The membership function is the graph representing the amplitude of each input participation. The rules use the input membership values as a reference to determine their impact on the final outputs [52]. In this work, the Gaussian type is chosen for all the nodes of the BN because it provides less errors in the prediction of the data compared to the others, notably the triangular and trapezoidal forms [53]. Fig. 3 shows an example of this function for the ‘routing’ variable. Fig. 4. Fuzzy inference of the ‘QoS in VANET’ variable. Fig. 3. Membership function for routing variable. The next step aggregates different conclusions of the Inference activated rules, as shown in Table V, and combines them into a In the remaining steps, we will assume that all CPs have been single value. This value is obtained by the union of all the generated and the method used will be explained for the node triggered conclusions, translated by the max operator. ‘QoS in VANET’. In this case, the inference mechanism aims TABLE V at calculating the table of CPs of this variable while considering THE ACTIVATED RULES OF THE “QOS IN VANET” all combinations of its parents: ‘routing’ and ‘quality of data’. Table IV presents the set of fuzzy rules used in this example. ACTIVATED LINGUISTIC VALUE FOR THE DEGREE OF RULES OUTPUT APPURTENANCE The open-source software Fispro is used to implement the fuzzy inference, which provides exact values from different R2 medium 0.135 subsets of the output variables. Fig. 4 presents the inference of R3 medium 0.607 the variable ‘QoS in VANET’ knowing that ‘data quality’ is R8 medium 0.135 medium and ‘routing’ is good, which corresponds to rule 3 in Table IV. R9 good 0.135 https://www.fispro.org/en/ Applied Computer Systems _________________________________________________________________________________________________2022/27 The ‘medium’ and ‘good’ values of the variable ‘QoS in improve QoS in VANET but can only consider a few VANET’ are 0.607 and 0.135, respectively. However, the parameters simultaneously. However, in reality, where lowest value of 0.001 is tolerated in the case where the decisions have to be made based on several criteria, these possibility is zero (here for the ‘bad value’) since each state of approaches are insufficient. Consequently, the presented the variable ‘QoS in VANET’ is possible, and this possibility approach is fruitful. It allows risk analysis at any time in the must be greater than zero. network and provides a massive database that can be extended QoS in VANET (bad) = 0.001 progressively using deep learning techniques. QoS in VANET (medium) = max (0.607, 0.135, 0.135) = 0.607 QoS in VANET (good) = 0.135 VI. CONCLUSION The sum of probabilities for each variable states must be One unknown problem can perturb the communication equal to 1. The CPs for the different states of the variable ‘QoS system or block the proper functioning of the network, causing in VANET’ of rule number 3 are calculated as follows: material and human losses. For that reason, equipping with risk P (QoS in VANET = bad | data quality = medium, routing = prediction tools is very important. This article proposes a FB high) = 0.001/ (0.001+0.607+0.135) = 0.001 model combining the BN approach that evaluates the causality P (QoS in VANET = medium | data quality = medium, relation between its nodes based on different data resources, routing = high) = 0.607/ (0.001+0.607+0.135) = 0.817 such as expert estimations or learning databases, and FL used P (QoS in VANET = good | data quality = medium, routing to generate all CPs needed. This model allows calculating the = high) = 0.135/ (0.001+0.607+0.135) = 0.182 QoS in VANET in the function of multiple criteria and provides By following this approach, all CPs for the variable ‘QoS in a database that helps determine the originality of risk degrading VANET’ for the different states of its antecedents were the QoS in the network. The complexity of the proposed computed as presented in Table VI. approach consists of the integration of all the factors that TABLE VI influence the QoS in VANET, where the absence of some variables can affect the efficacy of the prediction. CPS TABLE OF THE VARIABLE “QOS IN VANET” The future work will complete the actual model by QOS IN VANET generating the CPs of all nodes in the causal graph and propose QUALITY OF RULE ROUTING BAD MEDIUM GOOD some scenarios for proving the model efficacy. Furthermore, DATA some solutions will be presented in the case of bad QoS in the 1 medium 0.56 bad 0.17 0.994 0.005 0.001 network. 2 medium 0.56 Medium 0.57 0.002 0.931 0.067 3 medium 0.56 good 0.88 0.003 0.926 0.071 ACKNOWLEDGEMENT 4 bad 0.22 bad 0.17 0.995 0.004 0.001 The research has been supported by the Ministry of Higher 5 bad 0.22 medium 0.57 0.975 0.023 0.001 Education, Scientific Research and Innovation, the Digital Development Agency (DDA) and the National Centre for 6 bad 0.22 good 0.88 0.972 0.027 0.001 Scientific and Technical Research (CNRST) of Morocco 7 good 0.77 bad 0.17 0.995 0.004 0.001 (Smart DLSP Project – AL KHAWARIZMI IA-PROGRAM). 8 good 077 medium 0.57 0.002 0.935 0.063 9 good 0.77 good 0.88 0.001 0.035 0.964 REFERENCES [1] R. Bibi, Y. Saeed, A. Zeb, T. M. Ghazal, T. Rahman, R. A. Said, S. Abbas, V. DISCUSSION M. Ahmad, and M. A. Khan, “Edge AI-based automated detection and classification of road anomalies in VANET using deep learning,” The proposed FB model predicts the QoS state in VANET. Computational Intelligence and Neuroscience, vol. 2021, Art. no. By varying the values of the input parameters, their effects on 6262194, Sep. 2021. https://doi.org/10.1155/2021/6262194 [2] M. Gayathri and C. Gomathy, “A deep survey on types of cyber attacks the different nodes of the graph are estimated, implementing the in VANET,” J Crit Rev, vol. 8, no. 1, pp. 1029–1039, 2021. chrome- fuzzy inference method. extension://efaidnbmnnnibpcajpcglclefindmkaj/https://www.jcreview.co In many domains and studies, the combination of the BN and m/admin/Uploads/Files/61c07bf6658ea1.10579682.pdf [3] S. Lahdya and T. Mazri, “Security study of routing attacks in vehicular the FL system is used, for example, predicting the population ad-hoc networks (autonomous car),” The International Archives of status of a rare animal using the inference mechanism in the Photogrammetry, Remote Sensing and Spatial Information Sciences, ecological domain [54], determining the severity scale of a vol. 46, pp. 349–353, 2021. https://doi.org/10.5194/isprs-archives-XLVI- 4-W5-2021-349-2021 disease based on fuzzy medical rules [55], predicting the risks [4] A. E. C. Mondragon, C. S. Lalwani, E. S. C. Mondragon, and C. E. C. of delays in deliveries in logistic services [56] and others. Mondragon, “Facilitating multimodal logistics and enabling information To our knowledge, no prior work analyses risks in VANETs systems connectivity through wireless vehicular networks,” International Journal of Production Economics, vol. 122, no. 1, pp. 229–240, Nov. and predicts QoS, considering all previously mentioned 2009. https://doi.org/10.1016/j.ijpe.2009.05.023 parameters, which justifies the interest and originality of the [5] J. I. Katuka and M. S. A. Latiff, “VANETs and its related issues: An research presented in this article. In addition, this study extensive survey,” Journal of Theoretical & Applied Information provides an approach for generating massive data that help in Technology, vol. 66, no. 1, pp. 693–697, 2014. making decisions in the transportation domain. Some classical approaches and existing studies give optimal solutions to Applied Computer Systems _________________________________________________________________________________________________2022/27 [6] P. Pandey, “Effect of selfish behavior on network performance in [26] S. Tabar and A. Azadmanesh, “Quality of service in vehicular ad hoc VANET,” in Fifth International Conference on Communication Systems networks,” in IAENG Transactions on Engineering Sciences: Special and Network Technologies, Gwalior, India, Apr. 2015, pp. 693–697. Issue for the International Association of Engineers Conferences 2016, https://doi.org/10.1109/CSNT.2015.63 World Scientific, 2018, pp. 388–401. [7] S. Harrabi, I. B. Jaafar, and K. Ghedira, “Performance analysis of https://doi.org/10.1142/9789813230774_0028 VANETs routing protocols,” Res. Square, Univ. Mannouba, Manouba, [27] P. D. Dorge, S. S. Dorle, M. B. Chakole, and D. K. Thote, “Improvement Tunisia, Tech. Rep., 2021. of QoS in VANET with different mobility patterns,” in International [8] E. Singh, “Mobility and scalability management issues in VANET,” Int J Conference on Radar, Communication and Computing (ICRCC), Adv Res Sci Eng, vol. 7, pp. 151–155, 2018. Tiruvannamalai, India, Dec. 2012, pp. 206–209. [9] A. Ahamed and H. Vakilzadian, “Issues and challenges in VANET https://doi.org/10.1109/ICRCC.2012.6450578 routing protocols,” in Proc. IEEE Int. Conf. Electro/Inf. Technol. (EIT), [28] G. Parimala, S. Nageswararao, and K. LakshmiNadh, “Ddsrc: Algorithm Rochester, MI, USA, May 2018, pp. 723–728. for improving QoS in VANET,” International Journal of Recent https://doi.org/10.1109/EIT.2018.8500180 Technology and Engineering (IJRTE), vol. 7, no. 6S5, pp. 1327–1331, [10] P. Mundhe, S. Verma, and S. Venkatesan, “A comprehensive survey on Apr. 2019. authentication and privacy-preserving schemes in VANETs,” Computer [29] M. J. Sayadi, M. Fathy, and L. Mahaki, “Improving the quality of service Science Review, vol. 41, Art. no. 100411, Aug. 2021. in the VANET by detecting and removing unused messages,” https://doi.org/10.1016/j.cosrev.2021.100411 International Journal of Information & Communication Technology [11] J. Cui, D. Wu, J. Zhang, Y. Xu, and H. Zhong, “An efficient authentication Research, vol. 4, no. 3, pp. 107–112, 2012. scheme based on semi-trusted authority in VANETs,” IEEE Transactions [30] N. K. Kitson, A. C. Constantinou, Z. Guo, Y. Liu, and K. Chobtham, “A on Vehicular Technology, vol. 68, no. 3, pp. 2972–2986, Mar. 2019. survey of Bayesian network structure learning,” arXiv preprint, https://doi.org/10.1109/TVT.2019.2896018 arXiv:2109.11415, 2021. chrome- [12] A. Quyoom, A. A. Mir, and A. Sarwar, “Security attacks and challenges extension://efaidnbmnnnibpcajpcglclefindmkaj/https://arxiv.org/ftp/arxi of VANETs: a literature survey,” Journal of Multimedia Information v/papers/2109/2109.11415.pdf System, vol. 7, no. 1, pp. 45–54, Mar. 2020. [31] M. Scanagatta, A. Salmeron, and F. Stella, “A survey on Bayesian https://doi.org/10.33851/JMIS.2020.7.1.45 network structure learning from data,” Progress in Artificial Intelligence, [13] M. Arif, G. Wang, M. Z. A. Bhuiyan, T. Wang, and J. Chen, “A survey vol. 8, no. 4, pp. 425–439, May 2019. https://doi.org/10.1007/s13748- on security attacks in VANETs: Communication, applications and 019-00194-y challenges,” Vehicular Communications, vol. 19, Art. no. 100179, Oct. [32] N. R. Siddiqui, K. A. Khaliq, and J. Pannek, “VANET security analysis 2019. https://doi.org/10.1016/j.vehcom.2019.100179 on the basis of attacks in authentication,” in Dynamics in Logistics, [14] C. Sun, J. Liu, X. Xu, and J. Ma, “A privacy-preserving mutual Lecture Notes in Logistics, M. H. Kotzab and J. Pannek, Eds. Springer, authentication resisting dos attacks in VANETs,” IEEE Access, vol. 5, 2017, pp. 491–502. https://doi.org/10.1007/978-3-319-45117-6_43 pp. 24012–24022, 2017. https://doi.org/10.1109/ACCESS.2017.2768499 [33] K. B. Kelarestaghi, M. Foruhandeh, K. Heaslip, and R. Gerdes, “Survey [15] I. A. Sumra, H. B. Hasbullah, and J.-l. B. AbManan, “Attacks on security on vehicular ad hoc networks and its access technologies security goals (confidentiality, integrity, availability) in VANET: A survey,” in vulnerabilities and countermeasures,” arXiv preprint, arXiv:1903.01541, Vehicular Ad-Hoc Networks for Smart Cities, vol. 306, A. Laouiti, 2019. chrome- A. Qayyum, M. S. Mohamad, Eds. Springer, Singapore, 2015, pp. 51–61. extension://efaidnbmnnnibpcajpcglclefindmkaj/https://arxiv.org/ftp/arxi https://doi.org/10.1007/978-981-287-158-9_5 v/papers/1903/1903.01541.pdf [16] R. Kaur, T. P. Singh, and V. Khajuria, “Security issues in vehicular Ad- [34] M. B. Mansour, C. Salama, H. K. Mohamed, and S. A. Hammad, hoc network (VANET),” in 2018 2nd International conference on trends “VANET security and privacy-an overview,” International Journal of in Electronics and Informatics (ICOEI), Tirunelveli, India, May 2018, Network Security & Its Applications (IJNSA), vol. 10, no. 2, Mar. 2018. pp. 884–889. https://doi.org/10.1109/ICOEI.2018.8553852 https://doi.org/10.2139/ssrn.3290553 [17] M. Azees, P. Vijayakumar, and L. Jegatha Deborah, “Comprehensive [35] G.-U. Rehman, A. Ghani, S. Muhammad, M. Singh, and D. Singh, survey on security services in vehicular ad-hoc networks,” IET Intelligent “Selfishness in vehicular delay-tolerant networks: A review,” Sensors, Transport Systems, vol. 10, no. 6, pp. 379–388, Aug. 2016. vol. 20, no. 10, Art. no. 3000, 2020. https://doi.org/10.3390/s20103000 https://doi.org/10.1049/iet-its.2015.0072 [36] A. Reyes, C. Barrado, M. Lopez, and C. Excelente, “Vehicle density in [18] M. N. Mejri, J. Ben-Othman, and M. Hamdi, “Survey on VANET security VANET applications,” Journal of Ambient Intelligence and Smart challenges and possible cryptographic solutions,” Vehicular Environments, vol. 6, no. 4, pp. 469–481, 2014. Communications, vol. 1, no. 2, pp. 53–66, Apr. 2014. https://doi.org/10.3233/AIS-140271 https://doi.org/10.1016/j.vehcom.2014.05.001 [37] A. V. Ostroukh and H. Elhadi, “Comparative study of routing protocols [19] J. Contreras-Castillo, S. Zeadally, and J. A. G Ibañez, “Solving vehicular in vehicular ad-hoc networks (VANETs),” International Journal of ad hoc network challenges with big data solutions,” IET Networks, vol. 5, Advanced Studies (iJAS), vol. 4, no. 2, pp. 9–14, 2014. no. 4, pp. 81–84, Jul. 2016. https://doi.org/10.1049/iet-net.2016.0001 https://doi.org/10.12731/2227-930X-2014-2-2 [20] A. Mchergui, T. Moulahi, and S. Zeadally, “Survey on artificial [38] C. Tripp-Barba, M. A. Igartua, L. U. Aguiar, A. M. Mezher, A. Zaldívar- intelligence (AI) techniques for vehicular ad-hoc networks (VANETs),” Colado, and I. Guerin-Lassous, “Available bandwidth estimation in GPSR Vehicular Communications, vol. 34, Art. no. 100403, Apr. 2021. for VANETs,” in Proceedings of the third ACM international symposium https://doi.org/10.1016/j.vehcom.2021.100403 on Design and analysis of intelligent vehicular networks and applications, [21] R. Shrestha, R. Bajracharya, and S. Y. Nam, “Challenges of future Nov. 2013, pp. 1–8. https://doi.org/10.1145/2512921.2516961 VANET and cloud-based approaches,” Wireless Communications and [39] T. Takagi and M. Sugeno, “Fuzzy identification of systems and its Mobile Computing, vol. 2018, Art. no. 5603518, May 2018. applications to modeling and control,” IEEE transactions on systems, https://doi.org/10.1155/2018/5603518 man, and cybernetics, vol. SMC-15, no. 1, pp. 116–132, Jan.-Feb. 1985. [22] M. Pourkiani, S. Adabi, S. Jabbehdari, and A. Khademzadeh, “Improving https://doi.org/10.1109/TSMC.1985.6313399 QoS in VANETs: A survey,” Journal of Advances in Computer [40] K. A. Hafeez, L. Zhao, Z. Liao, and B. N.-W. Ma, “The optimal radio Engineering and Technology, vol. 3, no. 3, pp. 153–166, 2017. propagation model in VANET,” in 2009 Fourth International Conference [23] M. Pourkiani, S. Jabbehdari, and A. Khademzadeh, “A novel protocol on Systems and Networks Communications. Porto, Portugal, Sep. 2009, stack for improving QoS in vehicular networks,” International Journal of pp. 6–11. https://doi.org/10.1109/ICSNC.2009.101 Computer Science and Information Security, vol. 14, no. 6, pp. 358– 367, [41] F. Li, W. Chen, and Y. Shui, “Study on connectivity probability of June 2016. VANETs under adverse weather conditions at 5.9 Ghz,” IEEE Access, [24] H. Fatemidokht and M. K. Rafsanjani, “QMM-VANET: An efficient vol. 8, pp. 547–555, Dec. 2019. clustering algorithm based on QoS and monitoring of malicious vehicles https://doi.org/10.1109/ACCESS.2019.2962089 in vehicular ad hoc networks,” Journal of Systems and Software, vol. 165, [42] M. C. Surugiu and I. N. Stancel, “Monitoring and warning system of a Art. no. 110561, Jul. 2020. https://doi.org/10.1016/j.jss.2020.110561 route based on roadside sensors in VANET,” Procedia Manufacturing, [25] A. Srivastava, S. Verma, N. Jhanjhi, M. Talib, A. Malhotra et al., vol. 32, pp. 745–752, 2019. https://doi.org/10.1016/j.promfg.2019.02.281 “Analysis of quality of service in VANET,” IOP Conference Series: [43] S. Abdelhamid, H. S. Hassanein, and G. Takahara, “Vehicle as a resource Materials Science and Engineering, vol. 993, no. 1, Art. no. 012061, (VaaR),” IEEE Network, vol. 29, no. 1, pp. 12–17, Jan. 2015. 2020. https://doi.org/10.1088/1757-899X/993/1/012061 https://doi.org/10.1109/MNET.2015.7018198 Applied Computer Systems _________________________________________________________________________________________________2022/27 [44] N. Mohamed, S. Sahib, N. Suryana, and B. Hussin, “Understanding Hafida Khalfaoui obtained her B.Sc. in network congestion effects on performance-articles review,” Journal of Electronic and Telecommunication Engineering Theoretical and Applied Information Technology, vol. 92, no. 2, Art. no. and her M.Sc. in Telecommunication Systems 311, 2016. and Computer Networks from Sultan Moulay [45] R. Chen, Z. Sheng, Z. Zhong, M. Ni, D. G. Michelson, and V. C. Leung, Slimane University, Beni Mellal, Morocco, in “Analysis on connectivity performance for vehicular ad hoc networks 2017 and 2019, respectively. She is following her subjected to user behavior,” in 2015 International Wireless PhD in Mathematics and Computer Science at Communications and Mobile Computing Conference (IWCMC). Sultan Moulay Slimane University. Her research Dubrovnik, Croatia, Aug. 2015, pp. 26–31. interests include computer science and network https://doi.org/10.1109/IWCMC.2015.7288932 security. [46] R. K. Aswed and M. A. Abdala, “End-to-end delay enhancement with E-mail: hafidakhalfaoui1996@gmail.com aodv in VANET,” International Journal of Enhanced Research in ORCID iD: https://orcid.org/0000-0002-9408- Science, Technology & Engineering, vol. 3, no. 11, pp. 227–232, Nov. 1301 2014, [47] S. Renooij, “Probability elicitation for belief networks: issues to Abdellah Azmani received his PhD degree in consider,” The Knowledge Engineering Review, vol. 16, no. 3, pp. 255– Industrial Computing in Dynamic System 269, Feb. 2001. https://doi.org/10.1017/S0269888901000145 Modelling and Artificial Intelligent at the [48] H. Sattar, I. S. Bajwa, J. Muhammad, M. F. Mushtaq, R. Kazmi, University of Science and Technology of Lille in M. Akram, M. Ashraf, U. Shafi et al., “Smart wound hydration 1991. He worked as a Professor at the Ecole monitoring using biosensors and fuzzy inference system,” Wireless Centrale of Lille, France and at the Institute of Communications and Mobile Computing, vol. 2019, Art. no. 8059629, Computer and Industrial Engineering from Lens, 2019. https://doi.org/10.1155/2019/8059629 France. He is a Professor at the Faculty of [49] M.-D. Pop, O. Prostean, T.-M. David, and G. Prostean, “Hybrid solution Science and Technology of Tangier, Morocco. combining Kalman filtering with Takagi–Sugeno fuzzy inference system He is a member of the Laboratory of Informatics, for online car-following model calibration,” Sensors, vol. 20, no. 19, Art. System and Telecommunication (LIST) and he no. 5539, Sep. 2020. https://doi.org/10.3390/s20195539 created the Intelligent Automation team, which [50] E. H. Mamdani and S. Assilian, “An experiment in linguistic synthesis he coordinates. He has contributed to many with a fuzzy logic controller,” International Journal of Man-Machine theses, scientific research projects and he elaborates and produces many IT and Studies, vol. 7, no. 1, pp. 1–13, 1975. https://doi.org/10.1016/S0020- decision support solutions for public administration, business management, 7373(75)80002-2 marketing and logistics. [51] M. H. Rad and M. Abdolrazzagh-Nezhad, “Data cube clustering with E-mail: a.azmani@uae.ac.ma improved DBSCAN based on fuzzy logic and genetic algorithm: ORCID iD: https://orcid.org/0000-0003-4975-3807 Designing and improving data cube clustering,” Information Technology and Control, vol. 49, no. 1, pp. 127–143, Mar. 2020. Abderrazak Farchane received his B.Sc. in https://doi.org/10.5755/j01.itc.49.1.23780 Computer Science and Engineering in June 2001 [52] A. Omar, A. Aous, A. Ali, and S. S. Balasem, “Comparison between the and M.Sc. in Computer Science and effects of different types of membership functions on fuzzy logic Telecommunication from the University of controller performance,” International Journal of Emerging Engineering Mohammed V Agdal, Rabat, Morocco, in 2003. Research and Technology, vol. 76, pp. 76–83, 2015. He obtained his PhD in Computer Science and [53] S. N. Mandal, J. P. Choudhury, and S. B. Chaudhuri, “In search of suitable Engineering at ENSIAS, Rabat, Morocco. He is fuzzy membership function in prediction of time series data,” currently an Associate Professor of Computer International Journal of Computer Science Issues, vol. 9, no. 3, pp. 293– Science at the Polydisciplinary Faculty, Sultan 302, 2012. Moulay Slimane University, Morocco. His areas [54] K.-R. Liu, J.-Y. Kuo, K. Yeh, C.-W. Chen, H.-H. Liang, and Y.-H. Sun, of interest are information coding theory, “Using fuzzy logic to generate conditional probabilities in Bayesian belief cryptography, and security. networks: a case study of ecological assessment,” International Journal E-mail: a.farchane@gmail.com of Environmental Science and Technology, vol. 12, no. 3, pp. 871–884, Dec. 2015. https://doi.org/10.1007/s13762-013-0459-x Said Safi received his B.Sc. degree in Electronics [55] V. Zarikas, E. Papageorgiou, and P. Regner, “Bayesian network from Cadi Ayyad University, Marrakech, construction using a fuzzy rule based approach for medical decision Morocco, in 1995. He obtained his M.Sc. and support,” Expert Systems, vol. 32, no. 3, pp. 344–369, Jun. 2015. PhD from Chouaib Doukkali University and Cadi https://doi.org/10.1111/exsy.12089 Ayyad University in 1997 and 2002, [56] O. E. Bouhadi, M. Azmani, A. Azmani, and M. A. el ftouh, “Using a respectively. He is currently a Professor of fuzzy-Bayesian approach for predictive analysis of delivery delay risk,” Science at the Multidisciplinary Faculty, Sultan International Journal of Advanced Computer Science and Applications, Moulay Slimane University, Beni Mellal, vol. 13, no. 7, pp. 316–326, 2022. Morocco. His general interests span the areas of https://doi.org/10.14569/IJACSA.2022.0130740 communications and signal processing, estimation, time-series analysis and system identification. Safi has more than 160 publications. His research currently focuses on transmitter and receiver diversity techniques for single and multi-user fading communication channels and on broadband wireless communication systems. E-mail: safi.said@gmail.com ORCID iD: https://orcid.org/0000-0003-3390-9037 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Applied Computer Systems de Gruyter

Using a Fuzzy-Bayesian Approach for Predicting the QoS in VANET

Loading next page...
 
/lp/de-gruyter/using-a-fuzzy-bayesian-approach-for-predicting-the-qos-in-vanet-46c4wT8MUO

References

References for this paper are not available at this time. We will be adding them shortly, thank you for your patience.

Publisher
de Gruyter
Copyright
© 2022 Hafida Khalfaoui et al., published by Sciendo
ISSN
2255-8691
eISSN
2255-8691
DOI
10.2478/acss-2022-0011
Publisher site
See Article on Publisher Site

Abstract

Applied Computer Systems ISSN 2255-8691 (online) ISSN 2255-8683 (print) December 2022, vol. 27, no. 2, pp. 101–109 https://doi.org/10.2478/acss-2022-0011 https://content.sciendo.com Using a Fuzzy-Bayesian Approach for Predicting the QoS in VANET 1* 2 3 4 Hafida Khalfaoui , Abdellah Azmani , Abderrazak Farchane , Said Safi 1,3,4 LIMATI Laboratory, Sultan Moulay Slimane University, Beni Mellal, Morocco Intelligent Automation Laboratory, Abdelmalek Essaadi University, Tetouan, Morocco Abstract – There are considerable obstacles in the transport implement a tool for predicting problems that disrupt the proper sector of developing countries, including poor road conditions, functioning of the network. Therefore, this approach allows poor road maintenance and congestion. The dire impacts of these studying the influence of input parameters, including routing, challenges could be extremely damaging to both human lives and security and data processing, on the QoS in VANET. After that, the economies of the countries involved. Intelligent Transportation the risk can be resolved or even eliminated to safeguard system Systems (ITSs) integrate modern technologies into existing connectivity and avoid the problems of disruption in the transportation systems to monitor traffic. Adopting Vehicular Ad- hoc Network (VANET) into the road transport system is one of the transport, generating additional costs, distribution delays and most ITS developments demonstrating its benefits in reducing ultimately dissatisfaction. incidents, traffic congestion, fuel consumption, waiting times and The remainder of this article is structured as follows: pollution. However, this type of network is vulnerable to many Section II shows the challenges of VANET. Section III presents problems that can affect the availability of services. This article the literature related to the QoS in VANET. Section IV explains uses a Fuzzy Bayesian approach that combines Bayesian Networks the construction of the FB model. Section V discusses the (BN) and Fuzzy Logic (FL) for predicting the risks affecting the quality of service in VANET. The implementation of this model approach proposed. The final section concludes this work and can be used for different types of predictions in the networking presents some future areas of research. field and other research areas. II. CHALLENGES OF VANET Keywords – Bayesian network, fuzzy-Bayesian, fuzzy logic, prediction, quality of service, risk analysis, VANET. Significant problems appear when using ad-hoc communication between vehicles moving in a road I. INTRODUCTION environment. The key challenges are as follow: VANET is a Mobile Ad-hoc Network (MANET) subclass. A. Routing Problems Its objective is to create a communication system for Vehicle- The routing protocol that establishes and maintains routes to-Vehicle (V2V), Vehicle-to-Infrastructure (V2I) and from source to destination is the most challenging issue in Infrastructure to-Infrastructure (I2I) that increases awareness in VANET [5]. Since the infrastructure of VANET is limited, the the route and allows acquiring real-time traffic events, such as routing mechanism must be executed by the vehicles accidents, pre-crash warnings and emergency electronic brake themselves, and the intermediate route packets sent between lights [1], [2]. The mobile nodes are smart vehicles supplied distant vehicles [6]. This mechanism suffers from many issues with flexible computing resources containing computers, due to the unique features of VANET, such as the dynamic geolocation systems (GPS), radars, network devices and topology caused by high mobility of vehicles and variable various sensors [3]. These vehicles are interconnected and can network density. Furthermore, the cars may join and quit the have an interface that allows the driver to receive and send network rapidly, causing frequent path disruption. messages to various network entities. Furthermore, they Consequently, developing an efficient protocol that keeps a incorporate a positioning device such as GPS and sensors that route is difficult [7]. Besides, some nodes act selfishly and detect changed conditions, such as the state of the trailer’s doors refuse to participate in the routing to save their resources, such (open or closed), the temperature inside, violent movements as storage and processing time [6], which can degrade and harm and non-programmed stops [4]. These characteristics give high the performance of the network [7]. There are additional issues visibility and control of different operations and activities that regarding the VANET routing protocols. The most common occur within the transportation and help predict risks and make are: the scalability and limited bandwidth leading to congestion right decisions. and interference in the network [8], signal transmission This paper aims at measuring the quality of service (QoS) in attenuations caused by adverse weather and obstacles, also VANET by implementing a Fuzzy-Bayesian (FB) approach. security issues caused by malicious nodes that can launch The latter starts by analysing the risks of VANET in order to Corresponding author’s e-mail: hafidakhalfaoui1996@gmail.com ©2022 Hafida Khalfaoui, Abdellah Azmani, Abderrazak Farchane, Said Safi. This is an open access article licensed under the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0). 101 Applied Computer Systems _________________________________________________________________________________________________2022/27 diverse attacks, such as a black hole, eavesdropping, wormhole Availability and rushing attacks [9]. Availability is a critical factor that aims at protecting the life of users in VANET. It ensures that the network is operational B. Security Problems and useful information is always available [18]. The most The security aspect is an important pillar in VANET. If an famous availability threats are denial of service (DOS) attacks, attacker arrives to launch an attack, it can have a significant jamming attacks, malware attacks, greedy behaviour attacks, impact on network operations and services. Passive attacks, spamming attacks and so on. Their common goals are to use the such as listening to the communication and using the secret data bandwidth unnecessarily to cause voluntary collisions, decrease to execute malicious activities, or active attacks, such as communication quality between vehicles and thus interrupt removing, modifying, or repeating messages to disrupt the services or prohibit other nodes from using VANET support network, are both possible. It can also program traffic signals to and services [17]. cause traffic jams or block streets [10]. VANET presents several security issues because of its unique characteristics, C. Data Processing Problems such as the open network environment and dynamic network In recent years, the VANET environment has been topology [11]. The following are the most notable security generating massive amounts of data due to sensor and threats. communication technologies. These data are very important in improving productivity, travel comfort, transportation safety Authentication and modern society’s economic prosperity. As the amount of In VANET, there are two types of authentication: identity data generated by VANET grows exponentially, its complexity authentication, which determines the legitimacy of nodes, and and processing will present immense challenges due to the lack message authentication, which prevents fraudulent of intelligent systems that manage the large quantities of data transmission of messages in the network. Several existing generated. authentication algorithms have solved some security problems As a result, VANET requires complementary technologies to in VANETs, but their security still needs improvement. help optimize and efficiently process all data collected from Impersonation, identity (ID) disclosure attacks, spoofing disparate sources. Big data and artificial intelligence (AI) attacks, etc., are types of authentication attacks. They are provide competitive advantages in this context, allowing generated when malicious nodes can create false identities or organisations to make better decisions, while reducing reveal other node identities to act as legitimate vehicles and operational time and costs. However, these technologies present benefit from or reinforce incorrect data in the network. For significant challenges in terms of energy, node storage capacity example, an attacker may pretend to be an emergency vehicle and computation costs [19]. With the recent emergence of new in order to slow down other cars [12]. integrated architectures and access technologies, the Privacy Protection computation burden of IA algorithms and big data can be alleviated by migrating complex computations to external cloud Privacy protection has extended concern for VANET. servers [20]. This solution also presents some problems in terms Vehicles in contact with it typically share messages without any of time and security. When the scalability of nodes increases, privacy or security, which means the message is not encrypted the existing cloud computing can hardly satisfy the needs of end and contains vehicle ID, speed, location and other specific users. In general, uploading to and downloading from the cloud information that is usually linked to the driver’s identity. If consume time and energy [21]. these private data are open to attackers, this latter can use them to profile users or launch various attacks, such as masquerading III. RELATED WORK and impersonation attacks that can mislead other vehicles and endanger road safety [13], [14]. QoS measures the service degree of satisfaction as presented to the end user. In telecommunication networks, QoS aims at Confidentiality achieving better communication behaviour by routing the Confidentiality is a necessary security requirement for packets correctly and using resources optimally [22]. exchanging safety and non-safety messages between vehicles. Providing QoS support to VANET is an active research area. The message content should be secure and inaccessible to non- This network has particular features, such as high mobility, authenticated users. Confidentiality also prevents malicious scalability, a lack of centralized control unit, a hidden node activities, such as man-in-the-middle and traffic analysis issue, limited resources and an insecure medium, which make attacks [15], [16]. QoS provisioning very complex. In VANET, each application has specific QoS requirements; Integrity for example, a safety application should have a minimum End- Data integrity is one of the VANET security services that to-End Delay (E2ED) and packet loss because if a warning focus on conserving and ensuring the precision and accuracy of message does not arrive at a destination quickly, it will not help data during their transmission. There are numerous ways in prevent an accident [23]. which the data integrity service can be compromised. The most In the literature, many efforts and studies are dedicated to the common data integrity threats are replay, message tampering factors that affect the QoS in VANET and the majority are and illusion attacks [17]. looking into the routing aspect. Applied Computer Systems _________________________________________________________________________________________________2022/27 The authors of [24] proposed QMM-VANET to maintain E2ED of data transmission and, consequently, on the routing security and connectivity in the VANET. It is a clustering quality in the network? routing protocol that considers QoS requirements and mobility Fig. 1 represents a hierarchical arrangement, with both constraints. Firstly, this protocol calculates the QoS value of mobility and density are being ancestors and parents of E2ED; each vehicle and chooses a trustworthy one as the cluster head. analogical routing is its descendant and its child. Then, it selects gateways among neighbouring nodes and uses This study started with exhaustive research to identify the a gateway recovery algorithm to pick another gateway if link main challenges of a VANET network as follows: fails. In comparison to other protocols, this one demonstrated • Risks related to data routing influenced by the route its efficacy in terms of packet delivery ratio and E2ED in the conditions, such as mobility, density of nodes and highway scenario. security attacks. In [25] and [26], the authors conduct brief analyses on QoS, • Risks related to the security of the data circulated in the which include some types of routing protocols that help network due to the lack of robust mechanisms, such as minimize delay and increase overall QoS in VANET. The authentication, confidentiality, privacy and integrity. authors of [27] also improved various QoS parameters for • Risks related to data processing facilitating to make various scenario networks, such as delay, jitter, throughput and decisions in the different activities. Processing is affected packet loss ratio, by employing various routing algorithms with by the quality and quantity of data collected from the different mobility patterns, adaptive modulation, multiple-input different entities of the network. multiple-output (MIMO) and coding (AMC) techniques. This study also identified all the parameters that could In [28], the authors exploit Road Side Units (RSUs) to influence the occurrence degree of the above-mentioned risks. accumulate and transmit traffic data. They attempt to find the This identification led to conceiving the nature of the shortest route to the destination based on the cost of each dependencies between the different events and defining the transmission path using an effective algorithm called Dijkstra’s causal architecture of the Bayesian model. The following steps Dedicated Short Range Communication (DDSRC). are proposed for the analysis of QoS in VANET developed by According to the authors of [29], if the number of packets the BN: transmitted in the network increases, a queue of packets may be Step 1: Define the architecture of the BN (causal graph). created in the vehicle, and the response time of some packets Step 2: Generate conditional probabilities of intermediate may expire before being processed, causing a decreased service effects and final impacts. rate. They attempted to improve VANET performance by A. Architecture of the Bayesian Network of QoS in VANET removing useless or unused packets, in contrast to related works The architecture of the BN is based on the identification of that investigated the increasing service rate by changing the the relations between the nodes constituting the graph, which is parameters and properties of scheduler algorithms. divided according to their typology into three classes: In overall, it can be concluded that none of the previous QoS • The input parameters of the graph: these parameters solutions met all of the QoS criteria. represent the input nodes of the network. • Intermediate effects: the direct effects that are broken IV. MODELLING THE RISK OF QOS IN VANET USING A FUZZY- down into input parameters that are the origin of the BAYESIAN APPROACH causes. A Bayesian Network (BN) is a probabilistic graphical model • Final impacts: the factors that directly influence the QoS developed by Pearl to aid in reasoning under uncertainty [30]. in VANET. BN is defined by a pair (G, O) with G = V, A is a directed The investigation of the various factors that influence QoS in acyclic graph that encodes a joint probability distribution over VANETs aids in identifying input parameters, as described in a finite set of categorical variables V, and the arcs A represent Table I, and intermediate effects and final impacts, as shown in direct relationships between them [31]. In addition, a set of Table II. The causal graph in Fig. 2 is developed based on the parameters O defines the behaviour of each variable caused by two tables and after studying the nature of the causal its parent in the graph. dependencies between them. In this study, this tool allows for the prediction of problems that disturb the proper functioning of the network by studying B. Generation of Conditional Probabilities of Intermediate Effects and Final Impacts the influence of input parameters on such problems. For example, what is the impact of mobility and density on the After constructing the BN graph, the conditional probabilities (CPs) for each variable must be computed. There are numerous sources of probabilistic information available, such as databases containing road traffic and attacks. In this study, the BN considers a large number of parameters that make the existence of a complete database a difficult task. Unfortunately, the available databases are private and Fig. 1. Example of a causal relationship. insufficiently rich to permit a reliable examination of the https://www.unb.ca/cic/datasets/index.html Applied Computer Systems _________________________________________________________________________________________________2022/27 needed probabilities [47]. Therefore, FL is proposed for lies in its resemblance to human reasoning and natural generating these CPs systematically. It is an effective solution language. for handling imprecise data and linguistic problems. Its strength TABLE I DESCRIPTION OF THE INPUT PARAMETERS OF THE CAUSAL GRAPH CATEGORY OF VARIABLE NAME DESCRIPTION VARIABLES Security Authentication It represents the mechanism used to provide authenticity of both the data origin and also verify its sender. Integrity It represents the mechanism used to ensure that a sent message should reach the destination on the chosen path without any alteration [32]. Confidentiality The mechanism ensures that only authorized users can access the data. Failure to meet this requirement compromises the security of the exchanged data and the privacy of users [33]. Privacy This mechanism guarantees the anonymity of drivers and prevents an unauthorized party from knowing the combination between the sender’s real identity and private information related to traffic messages, such as vehicle identity, direction and location [34]. Inadvertence It refers to threats that result from trusted insiders who do not have malicious intent but fail to manage security responsibly. As a result, a malicious outsider is frequently able to use the privileges of the inattentive insider to carry out security problems. Selfish behaviour It refers to selfish nodes in the network that may refuse to cooperate in forwarding messages in order to increase their own resources [35]. Routing Mobility It refers to the movement of nodes communicating wirelessly and varies in function of their speed. Density It is the number of vehicles driving concurrently on the road [36]. The routing protocol is used to maintain links in the network to perform packet forwarding. In the VANET, Type of protocol there are many types of routing protocols. Choosing an effective one is very difficult because it depends on various parameters, such as mobility model, environment and others [37]. Bandwidth It is the maximum throughput at which packets can be transmitted between nodes without disrupting any existing network flow [38]. Frequency It is the number of occurrences of a repeating event per unit of time. If the frequency is high, we get short-range communication, which results in poor connectivity [39]. Obstacle Things that can diffract or block the signals in the road, such as buildings, trees, hills and bridges [40]. Weather Adverse weather conditions provide a signal transmission attenuation, like rainy or snowy days and dusty weather. Therefore, they will negatively influence the connectivity of high-speed data links in the vehicle communication range [41]. Data processing Internet Amount of data uploaded from the internet that can improve the productivity, travel comfort and modern society’s economic prosperity. OBU Amount of data exchanged with OBU. It concludes V2V and V2I communications. Sensors Amount of data collected from sensor systems related to state of vehicle or the external environment (based on radar systems, video cameras, image processing, etc.) [42]. Type of It is critical to choose the suitable type of data processing to avoid negative effect on the data output or the final processing product. Device resources These are resources/components that allow vehicle to act as a mobile service provider of sensing, data storage, data relaying, computing, infotainment and localization services [43]. TABLE II DESCRIPTION OF INTERMEDIATE EFFECTS AND FINAL IMPACTS INTERMEDIATE DESCRIPTION EFFECTS Signal transmission It represents the quality of signal used to transmit data between nodes. Interference Interference is caused by any signal other than the one configured in the network, and it obstructs the network normal operation. It usually causes slower speeds, higher latency, frequent disconnects, and sometimes, a complete failure to connect. Congestion This variable describes the state of the network when there is an overload of the links, which progressively affects the network performance with an increase in the transmission delay and a decrease in throughput [44]. Connectivity This property ensures successful data transmission from a source vehicle to a destination. It can be obtained if the number of active vehicles is increased in the network [45]. E2ED The time required to send a packet from its origin to its destination [46]. Packet loss The number of packets that fail to reach their destination. Applied Computer Systems _________________________________________________________________________________________________2022/27 Attacks It represents the number of attacks launched in the network. Security It represents the level of security in the network. Data processing Data processing is the process of collecting data and converting it into usable information (graphs, documents, etc.), to be interpreted by computers using various AI algorithms. This variable represents the quality of data processing used in the VANET. Quantity of data The amount of data generated in the network due to sensor and communication technologies. FINAL IMPACTS DESCRIPTION Routing This variable represents the quality of routing that has a great effect in the transmission of data between nodes. Quality of data Data quality is determined by many factors such as accuracy, completeness, confidentiality, validity, uniqueness and timeliness. QoS in VANET QoS measures the service degree of satisfaction as presented to the end user in the VANET. Fig. 2. BN modelling the analyses of risk of QoS in VANET. The implementation of the FL for generating CPs is done in TABLE III three principal steps [48], [49]: STATES OF THE BN NODES • Fuzzification: converts the input variables into a fuzzy VARIABLE LINGUISTIC VALUES subset using fuzzy linguistic values and membership functions. Authentication Robust, medium, weak • Inference: evaluates and combines the fuzzy rules to form Integrity Robust, medium, weak conclusions that give the fuzzy outputs of the system. Confidentiality Robust, medium, weak • Defuzzification: transforms the conclusions provided by Privacy Robust, medium, weak the inference engine into numeric values representing the Inadvertence High, medium, low final response of the fuzzy system. The most widely used inference methods are these of Attacks High, medium, low Mamdani [50] and Sugeno [39]. The main difference between Selfish behaviour High, medium, low them lies in the way the crisp output is obtained from the fuzzy Security Good, medium, bad inputs. Mamdani uses a defuzzification technique of fuzzy Frequency High, medium, low outputs, while Sugeno uses a weighted average to calculate the result values [51]. This study chooses the method of Sugeno Bandwidth Large, medium, small because it has a better processing time. As a result, generating Congestion High, medium, low CPs is done just with two first steps. Mobility High, medium, low Fuzzification Density High, medium, low To implement the fuzzification step, the fuzzy variables and Interferences High, medium, low their linguistic values are presented in Table III. Obstacle Many, medium, few Weather Good, medium, bad Applied Computer Systems _________________________________________________________________________________________________2022/27 TABLE IV Type of protocol Good, medium, bad STATES OF THE BN NODES Signal transmission Good, medium, bad RULE IF QUALITY OF AND THEN QOS IN Connectivity Good, medium, bad DATA ROUTING VANET E2ED Long, medium, short 1 medium bad bad Packet loss High, medium, low 2 medium medium medium OBU Large, medium, small 3 medium good medium Sensors Large, medium, small 4 bad bad bad Internet Large, medium, small 5 bad medium bad Quantity of data Large, medium, small 6 bad good bad Type of processing Good, medium, bad 7 good bad bad Devices resources High, medium, limited 8 good medium medium Data processing Good, medium, bad 9 good good good Quality of data Good, medium, bad Routing Good, medium, bad QoS in VANET Good, medium, bad The membership function is the graph representing the amplitude of each input participation. The rules use the input membership values as a reference to determine their impact on the final outputs [52]. In this work, the Gaussian type is chosen for all the nodes of the BN because it provides less errors in the prediction of the data compared to the others, notably the triangular and trapezoidal forms [53]. Fig. 3 shows an example of this function for the ‘routing’ variable. Fig. 4. Fuzzy inference of the ‘QoS in VANET’ variable. Fig. 3. Membership function for routing variable. The next step aggregates different conclusions of the Inference activated rules, as shown in Table V, and combines them into a In the remaining steps, we will assume that all CPs have been single value. This value is obtained by the union of all the generated and the method used will be explained for the node triggered conclusions, translated by the max operator. ‘QoS in VANET’. In this case, the inference mechanism aims TABLE V at calculating the table of CPs of this variable while considering THE ACTIVATED RULES OF THE “QOS IN VANET” all combinations of its parents: ‘routing’ and ‘quality of data’. Table IV presents the set of fuzzy rules used in this example. ACTIVATED LINGUISTIC VALUE FOR THE DEGREE OF RULES OUTPUT APPURTENANCE The open-source software Fispro is used to implement the fuzzy inference, which provides exact values from different R2 medium 0.135 subsets of the output variables. Fig. 4 presents the inference of R3 medium 0.607 the variable ‘QoS in VANET’ knowing that ‘data quality’ is R8 medium 0.135 medium and ‘routing’ is good, which corresponds to rule 3 in Table IV. R9 good 0.135 https://www.fispro.org/en/ Applied Computer Systems _________________________________________________________________________________________________2022/27 The ‘medium’ and ‘good’ values of the variable ‘QoS in improve QoS in VANET but can only consider a few VANET’ are 0.607 and 0.135, respectively. However, the parameters simultaneously. However, in reality, where lowest value of 0.001 is tolerated in the case where the decisions have to be made based on several criteria, these possibility is zero (here for the ‘bad value’) since each state of approaches are insufficient. Consequently, the presented the variable ‘QoS in VANET’ is possible, and this possibility approach is fruitful. It allows risk analysis at any time in the must be greater than zero. network and provides a massive database that can be extended QoS in VANET (bad) = 0.001 progressively using deep learning techniques. QoS in VANET (medium) = max (0.607, 0.135, 0.135) = 0.607 QoS in VANET (good) = 0.135 VI. CONCLUSION The sum of probabilities for each variable states must be One unknown problem can perturb the communication equal to 1. The CPs for the different states of the variable ‘QoS system or block the proper functioning of the network, causing in VANET’ of rule number 3 are calculated as follows: material and human losses. For that reason, equipping with risk P (QoS in VANET = bad | data quality = medium, routing = prediction tools is very important. This article proposes a FB high) = 0.001/ (0.001+0.607+0.135) = 0.001 model combining the BN approach that evaluates the causality P (QoS in VANET = medium | data quality = medium, relation between its nodes based on different data resources, routing = high) = 0.607/ (0.001+0.607+0.135) = 0.817 such as expert estimations or learning databases, and FL used P (QoS in VANET = good | data quality = medium, routing to generate all CPs needed. This model allows calculating the = high) = 0.135/ (0.001+0.607+0.135) = 0.182 QoS in VANET in the function of multiple criteria and provides By following this approach, all CPs for the variable ‘QoS in a database that helps determine the originality of risk degrading VANET’ for the different states of its antecedents were the QoS in the network. The complexity of the proposed computed as presented in Table VI. approach consists of the integration of all the factors that TABLE VI influence the QoS in VANET, where the absence of some variables can affect the efficacy of the prediction. CPS TABLE OF THE VARIABLE “QOS IN VANET” The future work will complete the actual model by QOS IN VANET generating the CPs of all nodes in the causal graph and propose QUALITY OF RULE ROUTING BAD MEDIUM GOOD some scenarios for proving the model efficacy. Furthermore, DATA some solutions will be presented in the case of bad QoS in the 1 medium 0.56 bad 0.17 0.994 0.005 0.001 network. 2 medium 0.56 Medium 0.57 0.002 0.931 0.067 3 medium 0.56 good 0.88 0.003 0.926 0.071 ACKNOWLEDGEMENT 4 bad 0.22 bad 0.17 0.995 0.004 0.001 The research has been supported by the Ministry of Higher 5 bad 0.22 medium 0.57 0.975 0.023 0.001 Education, Scientific Research and Innovation, the Digital Development Agency (DDA) and the National Centre for 6 bad 0.22 good 0.88 0.972 0.027 0.001 Scientific and Technical Research (CNRST) of Morocco 7 good 0.77 bad 0.17 0.995 0.004 0.001 (Smart DLSP Project – AL KHAWARIZMI IA-PROGRAM). 8 good 077 medium 0.57 0.002 0.935 0.063 9 good 0.77 good 0.88 0.001 0.035 0.964 REFERENCES [1] R. Bibi, Y. Saeed, A. Zeb, T. M. Ghazal, T. Rahman, R. A. Said, S. Abbas, V. DISCUSSION M. Ahmad, and M. A. Khan, “Edge AI-based automated detection and classification of road anomalies in VANET using deep learning,” The proposed FB model predicts the QoS state in VANET. Computational Intelligence and Neuroscience, vol. 2021, Art. no. By varying the values of the input parameters, their effects on 6262194, Sep. 2021. https://doi.org/10.1155/2021/6262194 [2] M. Gayathri and C. Gomathy, “A deep survey on types of cyber attacks the different nodes of the graph are estimated, implementing the in VANET,” J Crit Rev, vol. 8, no. 1, pp. 1029–1039, 2021. chrome- fuzzy inference method. extension://efaidnbmnnnibpcajpcglclefindmkaj/https://www.jcreview.co In many domains and studies, the combination of the BN and m/admin/Uploads/Files/61c07bf6658ea1.10579682.pdf [3] S. Lahdya and T. Mazri, “Security study of routing attacks in vehicular the FL system is used, for example, predicting the population ad-hoc networks (autonomous car),” The International Archives of status of a rare animal using the inference mechanism in the Photogrammetry, Remote Sensing and Spatial Information Sciences, ecological domain [54], determining the severity scale of a vol. 46, pp. 349–353, 2021. https://doi.org/10.5194/isprs-archives-XLVI- 4-W5-2021-349-2021 disease based on fuzzy medical rules [55], predicting the risks [4] A. E. C. Mondragon, C. S. Lalwani, E. S. C. Mondragon, and C. E. C. of delays in deliveries in logistic services [56] and others. Mondragon, “Facilitating multimodal logistics and enabling information To our knowledge, no prior work analyses risks in VANETs systems connectivity through wireless vehicular networks,” International Journal of Production Economics, vol. 122, no. 1, pp. 229–240, Nov. and predicts QoS, considering all previously mentioned 2009. https://doi.org/10.1016/j.ijpe.2009.05.023 parameters, which justifies the interest and originality of the [5] J. I. Katuka and M. S. A. Latiff, “VANETs and its related issues: An research presented in this article. In addition, this study extensive survey,” Journal of Theoretical & Applied Information provides an approach for generating massive data that help in Technology, vol. 66, no. 1, pp. 693–697, 2014. making decisions in the transportation domain. Some classical approaches and existing studies give optimal solutions to Applied Computer Systems _________________________________________________________________________________________________2022/27 [6] P. Pandey, “Effect of selfish behavior on network performance in [26] S. Tabar and A. Azadmanesh, “Quality of service in vehicular ad hoc VANET,” in Fifth International Conference on Communication Systems networks,” in IAENG Transactions on Engineering Sciences: Special and Network Technologies, Gwalior, India, Apr. 2015, pp. 693–697. Issue for the International Association of Engineers Conferences 2016, https://doi.org/10.1109/CSNT.2015.63 World Scientific, 2018, pp. 388–401. [7] S. Harrabi, I. B. Jaafar, and K. Ghedira, “Performance analysis of https://doi.org/10.1142/9789813230774_0028 VANETs routing protocols,” Res. Square, Univ. Mannouba, Manouba, [27] P. D. Dorge, S. S. Dorle, M. B. Chakole, and D. K. Thote, “Improvement Tunisia, Tech. Rep., 2021. of QoS in VANET with different mobility patterns,” in International [8] E. Singh, “Mobility and scalability management issues in VANET,” Int J Conference on Radar, Communication and Computing (ICRCC), Adv Res Sci Eng, vol. 7, pp. 151–155, 2018. Tiruvannamalai, India, Dec. 2012, pp. 206–209. [9] A. Ahamed and H. Vakilzadian, “Issues and challenges in VANET https://doi.org/10.1109/ICRCC.2012.6450578 routing protocols,” in Proc. IEEE Int. Conf. Electro/Inf. Technol. (EIT), [28] G. Parimala, S. Nageswararao, and K. LakshmiNadh, “Ddsrc: Algorithm Rochester, MI, USA, May 2018, pp. 723–728. for improving QoS in VANET,” International Journal of Recent https://doi.org/10.1109/EIT.2018.8500180 Technology and Engineering (IJRTE), vol. 7, no. 6S5, pp. 1327–1331, [10] P. Mundhe, S. Verma, and S. Venkatesan, “A comprehensive survey on Apr. 2019. authentication and privacy-preserving schemes in VANETs,” Computer [29] M. J. Sayadi, M. Fathy, and L. Mahaki, “Improving the quality of service Science Review, vol. 41, Art. no. 100411, Aug. 2021. in the VANET by detecting and removing unused messages,” https://doi.org/10.1016/j.cosrev.2021.100411 International Journal of Information & Communication Technology [11] J. Cui, D. Wu, J. Zhang, Y. Xu, and H. Zhong, “An efficient authentication Research, vol. 4, no. 3, pp. 107–112, 2012. scheme based on semi-trusted authority in VANETs,” IEEE Transactions [30] N. K. Kitson, A. C. Constantinou, Z. Guo, Y. Liu, and K. Chobtham, “A on Vehicular Technology, vol. 68, no. 3, pp. 2972–2986, Mar. 2019. survey of Bayesian network structure learning,” arXiv preprint, https://doi.org/10.1109/TVT.2019.2896018 arXiv:2109.11415, 2021. chrome- [12] A. Quyoom, A. A. Mir, and A. Sarwar, “Security attacks and challenges extension://efaidnbmnnnibpcajpcglclefindmkaj/https://arxiv.org/ftp/arxi of VANETs: a literature survey,” Journal of Multimedia Information v/papers/2109/2109.11415.pdf System, vol. 7, no. 1, pp. 45–54, Mar. 2020. [31] M. Scanagatta, A. Salmeron, and F. Stella, “A survey on Bayesian https://doi.org/10.33851/JMIS.2020.7.1.45 network structure learning from data,” Progress in Artificial Intelligence, [13] M. Arif, G. Wang, M. Z. A. Bhuiyan, T. Wang, and J. Chen, “A survey vol. 8, no. 4, pp. 425–439, May 2019. https://doi.org/10.1007/s13748- on security attacks in VANETs: Communication, applications and 019-00194-y challenges,” Vehicular Communications, vol. 19, Art. no. 100179, Oct. [32] N. R. Siddiqui, K. A. Khaliq, and J. Pannek, “VANET security analysis 2019. https://doi.org/10.1016/j.vehcom.2019.100179 on the basis of attacks in authentication,” in Dynamics in Logistics, [14] C. Sun, J. Liu, X. Xu, and J. Ma, “A privacy-preserving mutual Lecture Notes in Logistics, M. H. Kotzab and J. Pannek, Eds. Springer, authentication resisting dos attacks in VANETs,” IEEE Access, vol. 5, 2017, pp. 491–502. https://doi.org/10.1007/978-3-319-45117-6_43 pp. 24012–24022, 2017. https://doi.org/10.1109/ACCESS.2017.2768499 [33] K. B. Kelarestaghi, M. Foruhandeh, K. Heaslip, and R. Gerdes, “Survey [15] I. A. Sumra, H. B. Hasbullah, and J.-l. B. AbManan, “Attacks on security on vehicular ad hoc networks and its access technologies security goals (confidentiality, integrity, availability) in VANET: A survey,” in vulnerabilities and countermeasures,” arXiv preprint, arXiv:1903.01541, Vehicular Ad-Hoc Networks for Smart Cities, vol. 306, A. Laouiti, 2019. chrome- A. Qayyum, M. S. Mohamad, Eds. Springer, Singapore, 2015, pp. 51–61. extension://efaidnbmnnnibpcajpcglclefindmkaj/https://arxiv.org/ftp/arxi https://doi.org/10.1007/978-981-287-158-9_5 v/papers/1903/1903.01541.pdf [16] R. Kaur, T. P. Singh, and V. Khajuria, “Security issues in vehicular Ad- [34] M. B. Mansour, C. Salama, H. K. Mohamed, and S. A. Hammad, hoc network (VANET),” in 2018 2nd International conference on trends “VANET security and privacy-an overview,” International Journal of in Electronics and Informatics (ICOEI), Tirunelveli, India, May 2018, Network Security & Its Applications (IJNSA), vol. 10, no. 2, Mar. 2018. pp. 884–889. https://doi.org/10.1109/ICOEI.2018.8553852 https://doi.org/10.2139/ssrn.3290553 [17] M. Azees, P. Vijayakumar, and L. Jegatha Deborah, “Comprehensive [35] G.-U. Rehman, A. Ghani, S. Muhammad, M. Singh, and D. Singh, survey on security services in vehicular ad-hoc networks,” IET Intelligent “Selfishness in vehicular delay-tolerant networks: A review,” Sensors, Transport Systems, vol. 10, no. 6, pp. 379–388, Aug. 2016. vol. 20, no. 10, Art. no. 3000, 2020. https://doi.org/10.3390/s20103000 https://doi.org/10.1049/iet-its.2015.0072 [36] A. Reyes, C. Barrado, M. Lopez, and C. Excelente, “Vehicle density in [18] M. N. Mejri, J. Ben-Othman, and M. Hamdi, “Survey on VANET security VANET applications,” Journal of Ambient Intelligence and Smart challenges and possible cryptographic solutions,” Vehicular Environments, vol. 6, no. 4, pp. 469–481, 2014. Communications, vol. 1, no. 2, pp. 53–66, Apr. 2014. https://doi.org/10.3233/AIS-140271 https://doi.org/10.1016/j.vehcom.2014.05.001 [37] A. V. Ostroukh and H. Elhadi, “Comparative study of routing protocols [19] J. Contreras-Castillo, S. Zeadally, and J. A. G Ibañez, “Solving vehicular in vehicular ad-hoc networks (VANETs),” International Journal of ad hoc network challenges with big data solutions,” IET Networks, vol. 5, Advanced Studies (iJAS), vol. 4, no. 2, pp. 9–14, 2014. no. 4, pp. 81–84, Jul. 2016. https://doi.org/10.1049/iet-net.2016.0001 https://doi.org/10.12731/2227-930X-2014-2-2 [20] A. Mchergui, T. Moulahi, and S. Zeadally, “Survey on artificial [38] C. Tripp-Barba, M. A. Igartua, L. U. Aguiar, A. M. Mezher, A. Zaldívar- intelligence (AI) techniques for vehicular ad-hoc networks (VANETs),” Colado, and I. Guerin-Lassous, “Available bandwidth estimation in GPSR Vehicular Communications, vol. 34, Art. no. 100403, Apr. 2021. for VANETs,” in Proceedings of the third ACM international symposium https://doi.org/10.1016/j.vehcom.2021.100403 on Design and analysis of intelligent vehicular networks and applications, [21] R. Shrestha, R. Bajracharya, and S. Y. Nam, “Challenges of future Nov. 2013, pp. 1–8. https://doi.org/10.1145/2512921.2516961 VANET and cloud-based approaches,” Wireless Communications and [39] T. Takagi and M. Sugeno, “Fuzzy identification of systems and its Mobile Computing, vol. 2018, Art. no. 5603518, May 2018. applications to modeling and control,” IEEE transactions on systems, https://doi.org/10.1155/2018/5603518 man, and cybernetics, vol. SMC-15, no. 1, pp. 116–132, Jan.-Feb. 1985. [22] M. Pourkiani, S. Adabi, S. Jabbehdari, and A. Khademzadeh, “Improving https://doi.org/10.1109/TSMC.1985.6313399 QoS in VANETs: A survey,” Journal of Advances in Computer [40] K. A. Hafeez, L. Zhao, Z. Liao, and B. N.-W. Ma, “The optimal radio Engineering and Technology, vol. 3, no. 3, pp. 153–166, 2017. propagation model in VANET,” in 2009 Fourth International Conference [23] M. Pourkiani, S. Jabbehdari, and A. Khademzadeh, “A novel protocol on Systems and Networks Communications. Porto, Portugal, Sep. 2009, stack for improving QoS in vehicular networks,” International Journal of pp. 6–11. https://doi.org/10.1109/ICSNC.2009.101 Computer Science and Information Security, vol. 14, no. 6, pp. 358– 367, [41] F. Li, W. Chen, and Y. Shui, “Study on connectivity probability of June 2016. VANETs under adverse weather conditions at 5.9 Ghz,” IEEE Access, [24] H. Fatemidokht and M. K. Rafsanjani, “QMM-VANET: An efficient vol. 8, pp. 547–555, Dec. 2019. clustering algorithm based on QoS and monitoring of malicious vehicles https://doi.org/10.1109/ACCESS.2019.2962089 in vehicular ad hoc networks,” Journal of Systems and Software, vol. 165, [42] M. C. Surugiu and I. N. Stancel, “Monitoring and warning system of a Art. no. 110561, Jul. 2020. https://doi.org/10.1016/j.jss.2020.110561 route based on roadside sensors in VANET,” Procedia Manufacturing, [25] A. Srivastava, S. Verma, N. Jhanjhi, M. Talib, A. Malhotra et al., vol. 32, pp. 745–752, 2019. https://doi.org/10.1016/j.promfg.2019.02.281 “Analysis of quality of service in VANET,” IOP Conference Series: [43] S. Abdelhamid, H. S. Hassanein, and G. Takahara, “Vehicle as a resource Materials Science and Engineering, vol. 993, no. 1, Art. no. 012061, (VaaR),” IEEE Network, vol. 29, no. 1, pp. 12–17, Jan. 2015. 2020. https://doi.org/10.1088/1757-899X/993/1/012061 https://doi.org/10.1109/MNET.2015.7018198 Applied Computer Systems _________________________________________________________________________________________________2022/27 [44] N. Mohamed, S. Sahib, N. Suryana, and B. Hussin, “Understanding Hafida Khalfaoui obtained her B.Sc. in network congestion effects on performance-articles review,” Journal of Electronic and Telecommunication Engineering Theoretical and Applied Information Technology, vol. 92, no. 2, Art. no. and her M.Sc. in Telecommunication Systems 311, 2016. and Computer Networks from Sultan Moulay [45] R. Chen, Z. Sheng, Z. Zhong, M. Ni, D. G. Michelson, and V. C. Leung, Slimane University, Beni Mellal, Morocco, in “Analysis on connectivity performance for vehicular ad hoc networks 2017 and 2019, respectively. She is following her subjected to user behavior,” in 2015 International Wireless PhD in Mathematics and Computer Science at Communications and Mobile Computing Conference (IWCMC). Sultan Moulay Slimane University. Her research Dubrovnik, Croatia, Aug. 2015, pp. 26–31. interests include computer science and network https://doi.org/10.1109/IWCMC.2015.7288932 security. [46] R. K. Aswed and M. A. Abdala, “End-to-end delay enhancement with E-mail: hafidakhalfaoui1996@gmail.com aodv in VANET,” International Journal of Enhanced Research in ORCID iD: https://orcid.org/0000-0002-9408- Science, Technology & Engineering, vol. 3, no. 11, pp. 227–232, Nov. 1301 2014, [47] S. Renooij, “Probability elicitation for belief networks: issues to Abdellah Azmani received his PhD degree in consider,” The Knowledge Engineering Review, vol. 16, no. 3, pp. 255– Industrial Computing in Dynamic System 269, Feb. 2001. https://doi.org/10.1017/S0269888901000145 Modelling and Artificial Intelligent at the [48] H. Sattar, I. S. Bajwa, J. Muhammad, M. F. Mushtaq, R. Kazmi, University of Science and Technology of Lille in M. Akram, M. Ashraf, U. Shafi et al., “Smart wound hydration 1991. He worked as a Professor at the Ecole monitoring using biosensors and fuzzy inference system,” Wireless Centrale of Lille, France and at the Institute of Communications and Mobile Computing, vol. 2019, Art. no. 8059629, Computer and Industrial Engineering from Lens, 2019. https://doi.org/10.1155/2019/8059629 France. He is a Professor at the Faculty of [49] M.-D. Pop, O. Prostean, T.-M. David, and G. Prostean, “Hybrid solution Science and Technology of Tangier, Morocco. combining Kalman filtering with Takagi–Sugeno fuzzy inference system He is a member of the Laboratory of Informatics, for online car-following model calibration,” Sensors, vol. 20, no. 19, Art. System and Telecommunication (LIST) and he no. 5539, Sep. 2020. https://doi.org/10.3390/s20195539 created the Intelligent Automation team, which [50] E. H. Mamdani and S. Assilian, “An experiment in linguistic synthesis he coordinates. He has contributed to many with a fuzzy logic controller,” International Journal of Man-Machine theses, scientific research projects and he elaborates and produces many IT and Studies, vol. 7, no. 1, pp. 1–13, 1975. https://doi.org/10.1016/S0020- decision support solutions for public administration, business management, 7373(75)80002-2 marketing and logistics. [51] M. H. Rad and M. Abdolrazzagh-Nezhad, “Data cube clustering with E-mail: a.azmani@uae.ac.ma improved DBSCAN based on fuzzy logic and genetic algorithm: ORCID iD: https://orcid.org/0000-0003-4975-3807 Designing and improving data cube clustering,” Information Technology and Control, vol. 49, no. 1, pp. 127–143, Mar. 2020. Abderrazak Farchane received his B.Sc. in https://doi.org/10.5755/j01.itc.49.1.23780 Computer Science and Engineering in June 2001 [52] A. Omar, A. Aous, A. Ali, and S. S. Balasem, “Comparison between the and M.Sc. in Computer Science and effects of different types of membership functions on fuzzy logic Telecommunication from the University of controller performance,” International Journal of Emerging Engineering Mohammed V Agdal, Rabat, Morocco, in 2003. Research and Technology, vol. 76, pp. 76–83, 2015. He obtained his PhD in Computer Science and [53] S. N. Mandal, J. P. Choudhury, and S. B. Chaudhuri, “In search of suitable Engineering at ENSIAS, Rabat, Morocco. He is fuzzy membership function in prediction of time series data,” currently an Associate Professor of Computer International Journal of Computer Science Issues, vol. 9, no. 3, pp. 293– Science at the Polydisciplinary Faculty, Sultan 302, 2012. Moulay Slimane University, Morocco. His areas [54] K.-R. Liu, J.-Y. Kuo, K. Yeh, C.-W. Chen, H.-H. Liang, and Y.-H. Sun, of interest are information coding theory, “Using fuzzy logic to generate conditional probabilities in Bayesian belief cryptography, and security. networks: a case study of ecological assessment,” International Journal E-mail: a.farchane@gmail.com of Environmental Science and Technology, vol. 12, no. 3, pp. 871–884, Dec. 2015. https://doi.org/10.1007/s13762-013-0459-x Said Safi received his B.Sc. degree in Electronics [55] V. Zarikas, E. Papageorgiou, and P. Regner, “Bayesian network from Cadi Ayyad University, Marrakech, construction using a fuzzy rule based approach for medical decision Morocco, in 1995. He obtained his M.Sc. and support,” Expert Systems, vol. 32, no. 3, pp. 344–369, Jun. 2015. PhD from Chouaib Doukkali University and Cadi https://doi.org/10.1111/exsy.12089 Ayyad University in 1997 and 2002, [56] O. E. Bouhadi, M. Azmani, A. Azmani, and M. A. el ftouh, “Using a respectively. He is currently a Professor of fuzzy-Bayesian approach for predictive analysis of delivery delay risk,” Science at the Multidisciplinary Faculty, Sultan International Journal of Advanced Computer Science and Applications, Moulay Slimane University, Beni Mellal, vol. 13, no. 7, pp. 316–326, 2022. Morocco. His general interests span the areas of https://doi.org/10.14569/IJACSA.2022.0130740 communications and signal processing, estimation, time-series analysis and system identification. Safi has more than 160 publications. His research currently focuses on transmitter and receiver diversity techniques for single and multi-user fading communication channels and on broadband wireless communication systems. E-mail: safi.said@gmail.com ORCID iD: https://orcid.org/0000-0003-3390-9037

Journal

Applied Computer Systemsde Gruyter

Published: Dec 1, 2022

Keywords: Bayesian network; fuzzy-Bayesian; fuzzy logic; prediction; quality of service; risk analysis; VANET

There are no references for this article.