Perception-assisted superimposed CSI feedback for millimeter wave massive MIMO systems

The channel state information (CSI) feedback in frequency division duplex (FDD) mode-based millimeter wave (mmWave) massive multiple-input multiple-output (mMIMO) systems encounters significant feedback overhead due to the huge number of the base station (BS) antennas. To address this issue, we introduce the superimposed scheme into the CSI feedback. However, severe superimposed interference and channel estimation (CE) errors at the user equipment (UE) significantly degrade the recovery accuracy of downlink CSI. Inspired by perception-assisted communications, a perception-assisted superimposed CSI feedback method is proposed in this paper. In the proposed method, we employ the classic multiple signal classification (MUSIC) algorithm to extract the angle of arrival (AoA) of resolvable echo paths to form the perception information. With this perception information, a perception-assisted CSI recovery scheme is developed. This scheme comprises two phases, referred to as the preprocessing phase and the iterative recovery phase. By utilizing the perception information, the angular filtering matrices are constructed in the preprocessing phase to eliminate the angular components of non-path directions. In the iterative recovery phase, the perception information is further leveraged to iteratively suppress the superimposed interference and CE errors, thereby refining the recovery of the downlink CSI and uplink user data sequences (UL-US). Numerical simulation results demonstrate that the proposed method effectively improves the recovery accuracy of the downlink CSI and UL-US. Additionally, the proposed method exhibits its robustness against parameter variations.