Cooperative Reinforcement Learning Based Adaptive Resource Allocation in V2V Communication

Abstract

Platooning is one of the key applications of Intelligent Transportation System (ITS) for the smart cities. Various wireless technologies have been proposed for meeting the stringent requirements of platooning. 3GPP has initiated standardization work for LTE based V2V communication. It offers potential means to support transmission of safety critical messages among platoon vehicles with high reliability, security and ultra low latency. However, efficient resource allocation has been a challenge in LTE based networks. In this paper, we propose a Cooperative-Reinforcement Learning (C-RL) based resource selection algorithm for communication among connected vehicles utilizing LTE-Direct technology. The proposal outperforms the distributed resource selection scheme in terms of actual time required for Cooperative Awareness Messages (CAM) dissemination among vehicles forming the platoon and performance of other vehicular links sharing the similar Resource Blocks (RBs). Simulation results shows the efficacy of the proposed algorithm in terms of efficient resource utilization and faster dissemination of messages among the connected vehicles.