Vector Perturbation Precoding for MIMO Broadcast Channel with Quantized Channel Feedback

Abstract

In multiple-input multiple-output (MIMO) broadcast channels, the base station transmits information to multiple noncooperative users simultaneously. With perfect channel state information (CSI) at the base station, vector perturbation (VP) precoding technique achieves full diversity order. In this paper, we study the symbol error rate (SER) performance of such systems with quantized channel feedback. The channel quantization is modeled by following the rate-distortion theory. Based on minimum mean square error (MMSE) criterion, we study the optimal design of precoding matrix and perturbation vector. Equivalent relations in terms of both MMSE and SER between quantized and perfect channel feedback cases are established, based on which we investigate the relation between feedback load and SER performance.