Efficient precoder quantization for time-varying MIMO wireless channels

Closed Loop MIMO wireless systems achieve improved link performance by adaptively tuning the spatial precoding according to channel conditions. The information that is needed for adaptation is usually collected at the transmitter via a low-rate feedback channel from the receiver. Reduction of the bandwidth required by the feedback channel is a critical topic for forthcoming wireless systems. In this paper a feedback reduction technique that exploits correlation in time-varying MIMO channels is proposed. Differently from literature, the proposed scheme does not require statistical training of the optimized codebook. Furthermore, an estimate of the feedback load in case of MISO systems is evaluated in closed form. Numerical results confirm the analysis and show that the proposed scheme achieves significant feedback reduction for slowly-variable channels without degrading throughput.