Efficient Trust Assessment Methods With Security Consideration In Vehicular Ad-Hoc Networks

Vehicular ad hoc networks(VANETs)are considered promising candidate in the development of future Intelligent Transportation System(ITS)and Vehicular Internet of Things(VIoT).VANETs are originally extended from Mobile ad hoc networks(MANETs),but the special characteristics like high speed,random movement of vehicles,and lack of infrastructure make VANETs different from existing communication networks.Meanwhile,emerging trends in future ITS allow vehicles to have on-board sensors,Wi-Fi technologies,and digital computers,which might help in real-time information exchange among neighbor vehicles.However,considering infected sensors,computer malfunctioning,and malicious attacks,legitimate vehicles might inject fake information into the network.Security threats such as malicious attacks and fake news dissemination will affect the owner's travel plan and even cause traffic accidents.Therefore the driving motivation of our study is to efficiently evaluate the credibility of shared information sent among vehicles.This is challenging with VANETs due to its special characteristics as mentioned above.To this end,previously proposed security solutions in VANETs much emphasize on the public key exchange or complex cryptographic methods.These traditional security solutions become infeasible in critical cases like high mobility of vehicles,inside malicious attacks from authorized users,and unawareness about driving behavior.To fill this gap,trust is taken as a complement to cryptography and trust management methods are usually adopted.Therefore,our study emphasizes on less complex trust assessment methods to efficiently disseminate fair information among vehicles with security consideration as well.Following are some contribution to our dissertation.Vehicles on road may travel in an arbitrary fashion due to which network topology may be different.In this condition first,we have proposed Assess-Trust(AT)algorithm for one to one accurate user trust assessment among neighbor vehicles in multi-hop fashion.This algorithm using novel three valued subjective logic(3VSL)trust model applies trust propagation and fusion operations on each sub-user of VANETs,and finally combine them into final opinion.Also,unlike previous complex cryptographic methods,AT takes first-hand observation and recommendations from neighbors.A consensus is made using AT algorithm to judge the trustworthiness of the concerned vehicle.In order to validate AT algorithm,a numerical analysis is carried out.In this evaluation,we change the trust and prior uncertainty parameters values and note their impacts in simulation results to verify whether it works fine as said for multi-hop trust assessment among vehicles in VANETs.In addition,bridge network or non-series-parallel graphs are also analyzed to see their impact on final opinion.From simulation results,it is obvious that it may affect the final computed opinion,which is previously removed by network canonization.However,one major concern of AT algorithm is its time complexity,which increases for assessing one to one user trust in large scale VANETs.Second,to further solve the low scalability of AT algorithm in large-scale VANETs,we have proposed three-valued secure routing(3VSR)scheme,which is based on the widely used selforganizing Ad-hoc On-demand Distance Vector(AODV)routing,novel(3VSL),and our proposed TrustWalker(TW)algorithm as an efficient route search scheme.In addition,our proposed TW algorithm is an optimal route search technique in 3VSR,to solve the problem of massive trust assessment.TW searches the entire vehicular topology in Breadth First Search(BFS)manner.TW algorithm uses a matrix to represent the large scale VANET,storing the credibility of all users in the form of vectors.The vector is iteratively updated as TW algorithm "walks" through the entire network.In order to validate our proposed 3VSR scheme,experimental setup is carried out.Performance analysis is done under different malicious attacks and vehicles mobility settings,and compared with two other routing schemes i.e.,original AODV and Authenticated Anonymous Secure Routing(AASR).Also,another test is performed between AT and TW algorithms to prove that TW has better execution time as a route search technique.The experimental results shows that our proposed 3VSR combined with TW algorithm can effectively solve the massive trust assessment problem.However,high mobility of vehicles may affect digital services with delay,which needs modern edge technology to be integrated alongside with VANETs.Third,to further solve the problem of efficient data handover in high mobility VANETs and intelligent information dissemination,we have proposed Software Defined Networking(SDN)idea as modern edge technology in VANETs,also called software defined VANETs.The core concept behind this proposed idea is to use the control and data planes of SDN to perform related operations among vehicles more efficiently.As SDN are famous for their open-flow characteristics,so devices are more programmable in it.Data plane is related to forward/receive the relevant information generated by moving vehicles.Control plane in each region maintains social connections among vehicles,which are build based on their trust scores.We believe that using SDN architecture the performance can be optimized in term of efficient and intelligent information dissemination,vehicles data storage,and security provided by the central controller.In experimental analysis,we have considered packet loss and delay as an important metric for performance measure in software defined VANETs.Performance analysis is carried out to make comparison among our proposed scheme,software defined mobile networks(SDMA),and fast location-based handoff(FLBH).Simulation results validate that software defined VANETs can bring significant improvement in throughput and minimizing handover delay.