Reinforcement Learning based Integrated Sensing and Communication for Automotive MIMO Radar

Integrated sensing and communication (ISAC) is a promising technique in vehicular transportation thanks to its substantial gains in size, cost, power consumption, electromagnetic compatibility and spectrum congestion. In this paper, we propose a reinforcement learning (RL) based ISAC system with a multi-input-multi-output (MIMO) automotive radar. The target sensing and downlink communication are separately performed by dividing the transmit antennas into two non-overlapping but interweaving subarrays. We first design a RL framework to adaptively allocate the proper number of transmit antennas for the two subarrays under any unknown environment. The training is performed in the metrics of Cramer-Rao Bound (CRB) of direction of arrival (DOA) estimation for sensing and receive signal-to-noise (SNR) for communications, respectively. We proceed to propose a co-design method to jointly optimize the configurations of the two subarrays to further enhance the sensing accuracy with a constrained communication quality. The resultant problem is converted into the convex form via convex relaxation. Simulations are provided to demonstrate the adaptability and effectiveness of the proposed RL based ISAC system under the unkown environment.