Parallel IRS-Assisted Channel Estimation Based on Frequency Shifting

Intelligent reflecting surface (IRS) has recently emerged as a prominent communication paradigm due to its adaptability to wireless communication environments and low hardware cost. In IRS-assisted communication systems, channel state information (CSI) is essential for achieving a high beamforming gain. However, conventional IRS channel estimations suffer from the high complexity of IRS-related channels, increasing the estimation cost due to serial processing. In this paper, we propose a novel estimation method that leverages frequency shifts to improve channel estimation efficiency, which exploits the unique characteristics of IRS-related channels, utilizing multiple reflected signals as references. We also investigate the impact of individual channels' statistical properties and the reference signal's length on channel estimation performance. Furthermore, we introduce parallel and grouping estimation schemes for progressively element-wise channel estimation and accelerate the estimation process. Based on these, we design a corresponding beamforming scheme at IRS to improve the achievable data transmission rate. Extensive numerical results validate that the proposed estimation methods can replace pilot signals while achieving comparable accuracy and enabling simultaneous transmission and channel estimation with low computational load per time slot.