Feedback-Based Beam Training for Intelligent Reflecting Surface Aided mmWave Integrated Sensing and Communication
Xueyan Cao, Xiaoling Hu, Mugen Peng
Abstract
The intelligent reflecting surface (IRS) aided integrated sensing and communication (ISAC) is a promising technique in mmWave systems. However, most of the existing works on IRS-aided ISAC assume the availability of channel state information (CSI) along with high-overhead channel estimation. Besides, none has taken the impact of the echo interference from the communication user on the practical sensing performance into account. In this paper, we focus on the beamforming design in the presence of the destructive interference echo signal from the communication user and accordingly formulate a beam training problem for enhancing both sensing and communication performances. Considering that no CSI is available at the base station (BS), we propose a feedback-based joint active and passive beam training scheme by invoking a modified particle swarm optimization (PSO) method without requiring high-overhead cascaded channel estimation. Specifically, the BS updates the active and passive beams based on its received echo signal and user feedback information in the beam training period and then assigns them for more accurate sensing and more efficient communication in the ISAC period. Simulation results demonstrate the advantages of the proposed feedback-based beam training scheme over the conventional beamforming schemes without considering the interference echo signal, in terms of both sensing and communication performances. Moreover, it is shown that the destructive echo interference from the communication user can be effectively mitigated by limiting the communication signal-to-noise ratio, thereby achieving better sensing performance.