Research On Trusted Routing Protocol Based On Q Learning In Ad Hoc Networks

Mobile Ad Hoc networks are widely used in disaster relief,military operations,environmental monitoring and wilderness exploration because they can quickly establish networking and communications in the absence of network infrastructure.However,mobile Ad Hoc networks are facing many security threats due to frequent topology changes,wireless open links and limited node resources.Especially,each node of mobile Ad Hoc network acts as both host and router,transmitting data in multi-hop mode,so the routing security problems are very important.At present,the research on routing security of mobile Ad Hoc network is mainly focused on cryptography mechanism and trust evaluation mechanism at home and abroad.Although the scheme based on cryptography mechanism has high security,it can only solve static security problems,can't prevent malicious nodes from internal attacks and solve the link reliability problems caused by frequent changes in network topology.The scheme based on trust evaluation can establish and maintain dynamic trust relationship between nodes.It has the characteristics of resisting internal attacks,flexible deployment and application,and reducing security costs.So it is very suitable for mobile Ad Hoc network.At present,there are two problems in the routing security scheme of mobile Ad Hoc networks by applying trust evaluation mechanism.First,the existing trust model is difficult to accurately detect the attack of malicious nodes for effective defense.Second,in the routing stage,it is not possible to find the best route with the highest reliability from the whole network.Therefore,in this thesis a trust model with malicious behavior detection is designed,and Q learning is introduced to solve the problem of finding the most reliable optimal routing.The specific work is as follows: Firstly,aiming at the characteristics of mobile Ad Hoc network,a trust model based on weighted summation is established with direct trust attribute,attacking behavior attribute and cooperation attribute as trust factors.On this basis,a trust management module is designed.The module includes four modules: initialization of trust value,monitoring module,calculation and update module of trust value and malicious node management module.The behavior of Ad Hoc network node is advanced.Evaluate the line credibility,identify and isolate malicious nodes according to trust value,and do the basic work before routing.Secondly,the routing module is designed.Based on the AODV routing protocol and the proposed trust model,the AODV protocol is modified to generate multiple routes,and Q-learning is introduced to design the routing algorithm,which forms the TQAODV routing protocol based on trust model and Q-learning.Finally,the security and performance of TQAODV routing protocol is verified based on NS3 network simulation software.The experimental results show that the improved routing protocol can effectively identify and isolate malicious nodes,and find an optimal route with high reliability from the whole network.It significantly reduces the harm of malicious node attacks,improves the data packet throughput,and reduces the packet loss rate and end-to-end delay.It has very important theoretical significance and practical application value.