Reduced-Complexity Quasi-Optimum Detection for MIMO-OFDM Signals with Strong Nonlinear Distortion

Abstract

MIMO-OFDM (Multi-Input, Multi-Output Orthogonal Frequency Division Multiplexing) signals are very prone to nonlinear distortion effects, which can lead to high irreducible error floors. This can be particularly serious when low-resolution quantizers are employed. In this paper we consider MIMO-OFDM schemes with low- resolution quantizers. It is shown that, although this leads to strong nonlinear distortion effects and very poor performance when conventional receivers are employed, this is not necessarily the case when optimum receivers are considered. Since the complexity of the optimum receiver for MIMO-OFDM is prohibitively high, we develop sub- optimum receivers with moderate complexity. It is shown that these receivers can have huge performance gains when the nonlinear distortion levels are high, eliminating irreducible error floors and even outperforming linear MIMO-OFDM schemes in some scenarios.