Routing in VANETs: A fuzzy constraint Q-Learning approach

Abstract

Vehicular ad hoc networks (VANETs) can be used for the purpose of driving assistance, environment monitoring and entertainment. However, due to the vehicle movement, limited wireless resources and lossy feature of wireless channel, providing a reliable multi-hop communication in VANETs is particularly challenging. In this paper, we propose a VANET routing protocol which learns the optimal route by employing a fuzzy constraint Q-Learning algorithm. The protocol uses a fuzzy logic to evaluate a wireless link is whether good or not by considering multiple metrics of signal strength, available bandwidth and relative vehicle movement. Based on the evaluation of each wireless link, the proposed protocol learns the best route using the route request messages and hello messages. Upon reception of a route request message, each node maintains an evaluation value for each possible next hop node. In this way, the protocol can choose the best route, which is difficult to acquire in a typical reactive routing protocol. We show the effectiveness of the proposed protocol by using computer simulations.