A Q-learning Weighted Cluster Based Routing Protocol for VANETs
TL;DR
Simulation results obtained demonstrate that Q-WeCBR outperforms CBR, DSDV, and GPSR in terms of packet delivery ratio and throughput, confirming the effectiveness of clustering combined with learning-based routing for dynamic vehicular networks.
Abstract
Vehicular Ad-Hoc Networks are an emerging paradigm within Intelligent Transportation Systems (ITS), enabling communication between vehicles (V2V) and/or vehicles and infrastructure (V2I). These networks aim to enhance road safety and improve the driving experience. However, due to the high mobility of vehicles and frequent changes in their geographical positions, ensuring reliable data delivery remains a significant challenge. Clustering has emerged as a promising technique to improve scalability, reduce overhead, and enhance routing stability in VANETs. This paper first introduces a new classification framework for clustering-based routing protocols according to their operational and decision parameters. It then proposes the Q-learning Weighted ClusterBased Routing Protocol (Q-WeCBR), which combines clustering with reinforcement learning. Q-WeCBR improves cluster head (CH) selection through a weighted selection function and a maintenance phase that ensures cluster stability. In addition, the integration of Q-learning enables the protocol to adapt intelligently to topology changes by selecting the most reliable routes.Simulation results obtained with OMNeT++ and SUMO demonstrate that Q-WeCBR outperforms CBR, DSDV, and GPSR in terms of packet delivery ratio and throughput, confirming the effectiveness of clustering combined with learning-based routing for dynamic vehicular networks.