Research On Trust Evaluation And Detection Of Intrusion Nodes Based On Fuzzy Logic In VANETs

The arrival of 5G era marks that the development of Internet of vehicles has entered a new journey.To promote the development of the Internet of vehicles in the direction of automatic driving,traffic optimization management and intelligent vehicle system,people can enjoy the convenience brought by the progress of science and technology.Inevitably,the security issues derived from the Internet of vehicles(false broadcasting,intrusion attacks,privacy leakage,etc.)have become the research hotspot of scholars at home and abroad.Trust management mechanism can effectively evaluate the reliability of node behavior,identify the intrusion node efficiently and accurately,and remove the false messages from the network.In order to solve the huge communication storage and certificate management overhead brought by traditional cryptography and public key infrastructure solutions,as well as the limitations of intrusion detection system solutions with high deployment cost,which rely on data sets,the research scheme based on trust management is highly praised by many scholars for its advantages of light weight,high recognition rate,low complexity and low overhead,and most of the current trust management schemes are based on trust management Focus on the current behavior of the node,ignoring the past behavior of the node.Few trust studies can still get better evaluation performance in areas with low node density.Aiming at the above problems,through the modeling and analysis of different scenarios,this paper focuses on the trust evaluation and detection of intrusion nodes in the Internet of vehicles.The specific research contents are as follows:(1)Based on the shortcomings of the existing schemes,a decentralized trust evaluation model is built.The model evaluates the credibility of the target node through the collection of vehicle history and current social behavior and the suggestions of neighbor set,and introduces different types of intrusion attacks.Simulation experiments are carried out under different number of intrusion nodes to verify the effectiveness of the proposed model.(2)This paper proposes a direct trust evaluation scheme based on fuzzy logic,which considers three factors of quality index,kinetic energy index and cooperation index to achieve direct trust evaluation.This scheme can achieve higher accuracy of trust evaluation even in the case of fuzzy data or incomplete collection.(3)Based on the theory of collaborative filtering algorithm,the proposed level similarity of adjacent nodes is calculated to achieve the indirect trust evaluation of nodes.A group of experiments without indirect trust evaluation are set up to verify the effectiveness and necessity of indirect trust evaluation.