Research On The Trust Of On-Demand Routing Protocols In AD HOC Networks

As new technologies such as vehicle networking continue to progress, wireless ad hoc networks have achieved rapid development. In ad hoc networks, all nodes are able to send and forward packages, which can enable multi-hop transmission. Compared to traditional networks, ad hoc network is acentric, self-organized, multi-hop and flexible, and thus it is able to adapt to the military and vehicle networking. To accommodate multi-hop and mobile environments, people design a variety of routing protocols, including table-driven protocols and on-demand protocols, and on-demand routing protocol can adapt to the mobile environment better.In ad hoc networks, most routing protocols assume that each node in the network is cooperative and credible, which focus on finding good routing path, while ignoring the security and trust problems. Besides, because of many reasons such as the node mobility, unstable links will affect the network performance seriously. Thus, the presence of the malicious nodes and unstable links will cause adverse impact on the network.In this paper, considering the black-hole attack and the grey-hole attack, we build a novel trust model, and then a light-weight trust-based multipath routing protocol called LWT-AOMDV is proposed based on the proposed model, extending from the AOMDV. The model uses the successful delivery rate as the metric of trust evaluation, which is the direct trust value; every node can broadcast their own evaluation list of each neighbour nodes, which can be piggybacked by Hello packets through neighbour discovery mechanism, thereby the favourable rate of surrounding neighbours to the target node can be regarded as the indirect trust value; our trust evaluation model combines direct and indirect trust value as the comprehensive trust value. The current research for the trust solutions of the family of AODV routing protocols has two drawbacks:the trust assessment is imprecise and trust maintenance mechanism has relatively large overhead. The concept of the precursor list raised by us helps us achieve a precise multi-path trust assessment and our on-demand route maintenance mechanism in this protocol reduces control overhead by presenting the notion of path error instead of route error. Our simulation in NS2shows that our LWT-AOMDV protocol can significantly improve the delivery rate of the network in a variety of scenarios.In this paper, on account of the problem of the link stability, we propose a practical and efficient stability evaluation model based on the analysis of the received power, and on the base of AODV protocol, we design and implement a link stability-based routing protocol called S-AODV. The model uses the variance of the ratio of received powers as the direct evaluation of the stability, and we think of the average value of the direct stability of their neighbours as the indirect stability of the link. S-AODV protocol uses the link stability as the routing standards instead of hops and also resolves when the destination node sends a route reply packet and how the intermediate nodes deal with route request packets and other issues. In NS2simulation environment, we design multiple sets of experiments to determine the optimal parameter selection of the comprehensive link stability, and the experimental results show that our protocol can significantly reduce the frequency of route discovery at different speeds.