Resource Allocation in NR-V2X Mode 2 Using Multi Agent DQN

초록

V2X communication has been studied to increase road safety and traffic efficiency, and the most recently standardized technology is NR-V2X developed by third generation partnership project (3GPP). NR-V2X supports two modes of communication mode 1 and mode 2. In mode 2, vehicles reserve the resources based on their local observations using semi-persistent scheduling (SPS). In this method, if two or more vehicles select the same resource, a continuous resource collision occurs, and it makes communication performance greatly degraded. To overcome this, we propose a resource allocation method using multi-agent reinforcement learning (MARL). As agents, vehicles that transmit periodic cooperative awareness messages (CAM) are modeled. The state is composed of the received signals strength indicator (RSSI) that the agents received from each resource, and we set the total sum rate of all agents as the shared reward. The proposed method is compared with the random resource allocation in a highway scenario. The results show that the proposed method outperforms in terms of throughput performance. © 2023 IEEE.

키워드