Performance of MIMO Systems using Transmitter Preprocessing Based on Limited Noisy Feedback

Download

PDF Download (301Kb)

Description/Abstract

In this contribution we investigate the performance of Spatial Division Multiple Access (SDMA) multiple-input multiple-output (MIMO) systems using transmitter preprocessing, when the channel knowledge required for preprocessing is acquired by the receiver and conveyed to the transmitter via noise feedback channels that may also conflict fading. Specifically, in our system the MIMO channel impulse responses (CIRs) are vector quantized. Then, the CIR magnitudes and phases are conveyed to the transmitter via a feedback channel, which is noise contaminated and may also experience Rayleigh fading. At the transmitter, the CIRs used for transmit preprocessing are recovered using a soft estimator, which is optimum in the minimum mean-square error (MMSE) sense, and is implemented based on the so-called Hadamard soft-decoding principles. Our study and simulation results demonstrate that vector quantization combined with soft-decoding constitutes an efficient technique of feeding back the CIRs from the receiver to the transmitter. However, it is also known that the performance of the zero-forcing (ZF) or MMSE transmit preprocessing schemes is highly sensitive to the effect of quantization errors as well as to the feedback channel induced errors.