Full-Duplex Fluid-Antennas for Integrated Sensing and Backscatter Communication: A DRL-Based Fairness Design

Abstract

This letter investigates the efficacy of full-duplex fluid-antennas for a multi-target, multi-tag integrated sensing and backscatter communication system (ISABC) which exploits the extended degrees-of-freedom provided by fluid antennas at both the transmitter and the monostatic receiver. We formulate a max-min fairness-based joint sensing mutual information (MI) and communication rate maximization problem and develop a deep reinforcement learning (DRL)-based solution to effectively solve it. The proposed framework jointly optimizes the transmit precoder, radar signal, receive combiners, and the position vectors of the transmit and receive antennas. Simulation results show that the proposed scheme outperforms the benchmark system in both minimum communication rate and sensing MI, and approaches the performance of the alternating-optimization (AO) benchmark while offering a substantially lower runtime and better scalability.