Location-Sensing-Based Beamforming for Multi-IRS-Aided OFDM Communication Systems

Abstract

Intelligent reflecting surface (IRS)-aided orthogonal frequency division multiplexing communication has great potential to enhance system performance. Still, it faces the high overhead of acquiring cascaded channel state information. To address this, we propose a location-sensing-based beamforming scheme that avoids cascaded channel estimation, thereby reducing system overhead while maintaining communication performance. Specifically, a three-phase data transmission protocol is designed, comprising a location-information-aided beam-training phase, an equivalent-channel estimation phase, and a data transmission phase. The practical channel angle information for all IRS-associated links can be obtained from the sensed location information. Based on this, a bisection-search-based passive beam training method, an equivalent-channel estimation method, and an active beamforming method are proposed, corresponding to three phases that maximize the system communication rate. Finally, the advantages of the proposed location-sensing-based beamforming scheme and the effects of various system parameters are investigated. Simulation results suggest that the beamforming scheme can effectively obtain the base station transmit and IRS-reflection beams without the need for complicated cascaded channel estimation. Under the same achievable system rate, the proposed scheme reduces the channel estimation cost by 40%.