Research On Trust Management And Evaluation Methods In Internet Of Vehicles

The Internet of Vehicles is an important part of Intelligent Transportation System(ITS),which connects vehicles through the network and realizes the information exchange and interconnection between vehicles.The security and privacy issues of the Internet of vehicles are very important for drivers and vehicles.However,the Internet of Vehicles has the characteristics of dynamic changes in network structure and complex and diverse communication scenes,which makes the nodes in the Internet of vehicles need to solve more complex security and privacy protection problems than the traditional network nodes.Traditional security methods such as Public Key Infrastructure(PKI),digital certificates,encryption algorithms and access control can only solve the attacks from the outside of the system,but the internal attacks cannot be predicted and solved,and the trust management of the Internet of vehicles can also solve the internal attacks.Therefore,trust management is introduced into the Internet of vehicles to build the trust relationship between vehicles and the traffic system,so as to ensure the safety of vehicles and drivers.This paper has obtained the following research results:(1)This paper proposes a hierarchical trust management structure,which divides trust management tasks into different levels according to different entity roles in the Internet of Vehicles and simplifies the trust management process.The hierarchical Trust Management structure is divided into three layers:the first layer is the Domain Trust Management(DTM).The main entity role is the vehicle,which is responsible for collecting information about the interaction between vehicles.The second layer is Cluster Trust Management(CTM).The main entity role is RoadSide Unit(RSU),which is responsible for recommended trust value filtering and storage.The third layer is the Global Trust Management(GTM).The main entity role is the Trust Authority(TA),which is responsible for the calculation and dissemination of trust values.The structure illustrates the correlation and difference between each layer,which can solve the problems of data collection,storage and dissemination in the process of trust management.(2)A Recommend-value Filtering Trust Model(RFTM)was designed to calculate the trust values of vehicles in network of vehicles,which were divided into three aspects:direct trust,recommended trust,and comprehensive trust.In this model,the vehicle direct trust is calculated based on Bayesian inference method.Considering the timeliness of trust and trust switch attack,the time attenuation function,penalty factor and sliding window are introduced to calculate the direct trust value.The recommendation trust is then calculated.Because there will be malicious recommendation nodes in the calculation of exaggerating and denigrating behavior,so as to provide the wrong recommendation evaluation,a Fuzzy C-Means based recommendation filtering algorithm is designed,which is used to identify the wrong recommendation value,and then two factor similarity and reliability are introduced to determine the weight of the recommendation value.Finally,in the calculation of comprehensive trust value,the weighted method is used to combine direct trust and recommendation trust to form the final trust value.In this paper,adaptive weights are used to adapt to the dynamic changing environment of the Internet of vehicles,making trust evaluation more accurate.In order to verify the effectiveness of the proposed model,OMNeT++5.4.1,Veins4.7.1 and SUMOO.0.0 are used for joint simulation experiments.The trust model RFTM designed in this paper is compared with the two newer models NRB and NTM in the trust model.The results show that the proposed model can quickly and accurately calculate the trust value of vehicle nodes and has the ability to resist trust attacks.Compared with NRB and NTM,the accuracy of the proposed model is improved by 8.14%and 13.43%,respectively,with faster detection rate and higher convergence speed.It can reduce the impact of malicious nodes on network performance.