Research On Security Routing Technology In Vehicle Ad Hoc Network Based On Reinforcement Learning

With the development of intelligent transportation systems in recent years,Vehicular ad-hoc Network(VANET)has become a hot topic in the field of wireless Ad hoc networks.VANET mainly obtains rich road information through the communication between vehicles or between vehicles and infrastructure,so as to improve traffic efficiency and driving safety.However,due to its wireless communication mode,the open network makes the nodes vulnerable to attacks,which makes the current routing protocol face many security threats.In particular,due to the limited communication range of vehicles,data transmission in a multi-hop mode is required.Therefore,the routing security problem is very important.At present,the security of VANET routing is studied mainly from two aspects: cryptography mechanism and trust evaluation mechanism.Although the scheme based on cryptography mechanism is of high security,it has little effect on the bad behavior of authorized nodes inside the network,and cannot detect or even prevent the internal attack of malicious nodes,nor can it solve the link reliability problem caused by frequent changes in network topology.However,the scheme based on trust assessment can effectively resist the internal attack of VANET.By establishing and maintaining the dynamic trust relationship between entities,the nodes with bad behaviors are excluded,so as to ensure the reliability of communication.This thesis focuses on the VANET’s networking strategy and secure routing mechanism.Obviously,there are two main problems in the cluster-based routing protocol currently: one is that the security of network nodes is ignored so that the malicious attacks cannot be resisted effectively;The other is that the most reliable route can not be found from the whole network in the routing stage.Therefore,a trust-based clustering algorithm is designed in this thesis,and Q learning is introduced into the routing mechanism to solve the problem of finding the most reliable routing.The specific work is as follows:Firstly,according to the characteristics of VANET,the cluster-based networking strategy is researched.By introducing the two reference factors,that is relative mobility and centrality,a new cluster-head capability parameter is defined,and on this basis,a cluster-head election algorithm and a clustering mechanism are designed.At the same time,in order to guarantee the security and stability of the cluster structure,the trust model is introduced into the cluster maintenance mechanism to carry out trust accessment and maintenance for the cluster heads and cluster members,so as to improve the security and reliability of all participating nodes in the communication process,which provides a strong guarantee for the selection of safe routes.Secondly,the classical CBRP routing algorithm is improved by introducing Q-learning,and a security routing protocol TQCBRP is designed.In the proposed protocol,the trust of cluster nodes neighbors is evaluated through Q-learning,and the cluster head node with the maximum Q-value in one jump range,namely the highest credibility,is selected as the next jump to guide the routing forwarding node selection,establishing a stable and reliable routing forwarding path after multiple iterations.This method gives full consideration to the influence of all the cluster head nodes on routing forward,and realizes the global optimization.Finally,the security and performance of the proposed TQCBRP protocol are simulated by using the NS2 simulation tool.The experimental results show that the TQCBRP protocol can identify and exclude malicious nodes effectively,reduce the harm of malicious node attacks significantly.The proposed protocol is secure and valid.