Beam Splitting Sensing for Terahertz Near-Field Channel Estimation

Terahertz technology is considered as one of the key techniques for future 6G due to its ultra-wide bandwidth and ultra-high communication rate. However, terahertz communication channel estimation faces several significant challenges, one is that extremely-large-scale MIMO communication typically occurs in the near field, which will result in traditional far-field channel estimation schemes no longer being suitable. The other is the beam splitting effect caused by terahertz wideband communication, which leads to a large impact on the steering vector caused by different subcarrier frequencies. In order to solve these problems, this paper first studied the model of THz near-field beam splitting; secondly discussed the construction method of the THz near-field polar dictionary matrix; and then proposed a two-stage beam splitting sensing (Two-stage-BSS) algorithm to improve the accuracy of near-field channel estimation and reduce the impact of beam splitting. The experimental results validate the effectiveness of the proposed algorithm.