Optimal Resource Allocation for Integrated Sensing and Communications in Internet of Vehicles: A Deep Reinforcement Learning Approach

Abstract

Integrated sensing and communications (ISAC) technology is identified as a breakthrough in optimizing resource allocation and pursuing mutual benefits between radar sensing and wireless communications. In the complex Internet of Vehicles, however, the rapid increase in vehicles leads to a severe scarcity of spectrum resources that is fundamental for wireless communications. To fully exploit ISAC technology potential under strict resource constraints, this paper develops a joint channel, power, and bandwidth allocation scheme based on deep reinforcement learning to improve resource allocation efficiency. Firstly, a deep deterministic policy gradient algorithm is designed to ensure communication reliability by dynamically allocating wireless channels and transmitting power to vehicles based on their geographic positions. Then, we exploit a deep Q-network algorithm to maximize the communication rates by flexibly allocating frequency bandwidth. Simulation results demonstrate that the proposed algorithm exhibits higher spectral efficiency and communication rates than benchmark schemes.