Sensing Assisted Channel Extrapolation for TDD Massive MIMO-OFDM Systems

Yongbo Xiao,An Liu

DOI: https://doi.org/10.1109/vtc2024-spring62846.2024.10683017

2024-01-01

Abstract:The integration of radar sensing into communication networks for supporting various applications is an emerging area of interest in the field of integrated sensing and communication. In this paper, we propose a two-stage sensing-assisted channel extrapolation scheme for time-division duplex (TDD) massive MIMO-OFDM systems, utilizing radar-sensed scattering environment information (SEI) to enhance channel extrapolation performance. In Stage 1, user transmits uplink pilots (UPs) on a bandwidth part (BWP) that occupies only a fraction of the system bandwidth, for the base station (BS) to estimate the angle parameters through the proposed space-time multiple signal classification (ST-MUSIC) algorithm. However, it is difficult to achieve accurate channel extrapolation on the other BWPs because the delay resolution of a single BWP is limited. In Stage 2, the BS transmits full-band downlink pilots (DPs) aligned with the angles estimated in stage 1, and then performs joint scattering environment sensing and channel extrapolation based on the received UPs and reflected DPs. The EM-Turbo-CS algorithm is proposed for Stage 2 to exploit the SEI extracted from the full-band reflected DPs to dramatically improve the delay estimation accuracy, resulting in high-accuracy channel extrapolation. Meanwhile, a coarse estimation using the ST-MUSIC algorithm on the received signals helps reduce the complexity of the EM-Turbo-CS algorithm. Simulation results demonstrate that the proposed solution significantly improves the channel extrapolation performance.