Decentralized Joint Precoding With Pilot-Aided Beamformer Estimation

Abstract

Downlink beamforming techniques with low beamformer training overhead are proposed for joint processing (JP) coordinated multipoint transmission (CoMP). The objective is to maximize the weighted sum rate within joint transmission clusters without centralized beamformer processing, while accounting for uncertainty in the underlying channels. The proposed methods use time-division duplexing and pilot-based training with, possibly, nonorthogonal pilot sequences. The beamformer training is done without the explicit channel state information estimation, which greatly improves the robustness to pilot contamination. Best response and gradient-based decentralized algorithms are proposed and provide a tradeoff between computational complexity and fast convergence rate. The impact of feedback/backhaul quantization is also considered. The results show that JP CoMP is feasible with slow fading conditions and limited backhaul capacity by employing decentralized beamformer processing.