An Intelligent Machine Learning Based Routing Scheme for VANET

Currently, vehicular ad hoc networks (VANETs) are experiencing an explosive network traffic growth due to the rapid development of video streaming service, which causes large transmission delay and waste of bandwidth. However, the existing routing and caching strategies in VANETs are not efficient enough to cope with the time-varying network conditions, such as dynamic network topology. On the other hand, Knowledge Centric Network (KCN) is a new network architecture to solve above problems, which uses deep learning to manage network in a more intelligent and autonomous way. In this paper, we propose an Intelligent Routing Algorithm (IRA) based on deep belief network to provide high-quality multimedia service in VANETs, which considers a concept called user class. Deep belief network is used to extract feature vectors from users request sequence to divide user classes, which can improve quality of experience. We first divide mobile users into different user classes and users can discover requested video content from other users who belong to the same user class. Simulation results show that IRA achieves better network performance that other state-of-art routing algorithms in terms of delay in finding data and cache hit ratio.