Near-Optimal Performance With Low-Complexity ML-Based Detector for MIMO Spatial Multiplexing

Abstract

In Spatial Multiplexing MIMO systems, many powerful non-linear detection techniques as sphere decoding have emerged to overcome the performance limitations of linear detection techniques. However, these non-linear techniques suffer from high complexity that increases dramatically with the number of antennas and the modulation order. Hence, they cannot be implemented on highly parallel hardware architecture and are thus not suitable for real-time high data rate transmission. In this letter, a new detection technique is proposed to approach the optimal performance obtained by Maximum Likelihood (ML) detector without increasing the complexity significantly. This detector is denoted by OSIC-ML since it combines two techniques: the Ordered Successive Interference Cancellation (OSIC) and the ML. The proposed OSIC-ML detector shows a near-optimal performance at very low complexity even with large scale MIMO and imperfect channel estimation, where this complexity can be efficiently controlled to achieve the desired complexity-performance tradeoff.