A MAC Multi-channel Scheme Based on Learning-Automata for Clustered VANETs

Abstract

One of the main challenging issues of Vehicular Ad-Hoc Networks (VANETs) is the design of an efficient multi-channel Medium Access Control (MAC). Achieving efficient high throughput for Non-Safety services while maintaining bounded delay for time-critical road Safety applications is still a matter of investigation. In this paper, we propose a MAC Multi-channel Scheme based on Learning-automata for Clustered VANETs (LMMC). Our proposal relies on clustering approach, using single radio transceiver. Addressing the spectrum scarcity problem, the Cluster Head monitors the intra-cluster transmissions within the cluster according to a smart learning-automata model. The advantage of learning automatons is that the Cluster Head learns the traffic parameters of its cluster members without complication. Consequently, each cluster member is optimally assigned a fraction of TDMA slots proportional to its needs in terms of data transmissions. The major contributions of our LMMC protocol are: i) Optimal channel utilization while exchanging Safety or Non-Safety messages within a cluster. ii) Enhanced logical 100 ms MAC frame structure in a way that ensures bounded end-to-end delay of Safety applications. iii) Maximized throughput for throughput-sensitive transmissions.