Multi-Stage Precoder Design for Cooperative Massive MIMO Networks with Limited Feedback

Abstract

A multi-stage precoder can considerably lessen feedback overhead for channel state information (CSI) acquisition and backhaul overhead for CSI sharing in cooperative massive MIMO networks. In this paper, by devising precoder-adaptive subspace codebook, we propose dimension-optimized multi-stage precoder design to improve average sum rate. By taking into account all the interactions among the desired signal, same-group interference, inter-group interference, other-cell interference, and quantization error in the precoder design, we optimize the dimensions of the multi-stage precoder along with optimizing on selecting the dominant eigen-modes. Our simulation result verifies that our proposed design significantly increases the average sum rate compared to existing schemes.