Sensing Information-Assisted CSI Feedback

Abstract

Due to the limited transmission resources, as well as the limited computing resources and battery energy of user equipment (UE), the estimated downlink channel state information (CSI) is typically off-grid, deteriorating the channel sparsity and recovery accuracy of the downlink CSI at the base station (BS). To address these issues, we propose a sensing information-assisted CSI feedback method to recover the off-grid downlink CSI at the BS. Specifically, the proposed method utilizes the sensing information from distinct data modality to form the off-grid prior information. Subsequently, by leveraging the off-grid prior information and the received CSI, the complex path gain of the downlink CSI is recovered. Finally, a high resolution grid is constructed to recover the off-grid CSI matrix, thereby enhancing the recovery accuracy. Simulation results demonstrate that the proposed method effectively mitigates the off-grid effect of the downlink CSI at the BS. Furthermore, the proposed method presents performance improvement and robustness against parameter variations when compared to other methods.