A hybrid multiuser detection technique based on interference cancellation knowledge in MIMO communications

Abstract

In this paper, we propose a hybrid multiuser detection (MUD) technique, which is similar to the maximum likelihood detector (MLD) but with reduced complexity, for the application of multiple-input multiple-output (MIMO) communications. The proposed MUD technique avoids the exhaustive search required for the elimination of inter-antenna interference (IAI). Instead, the proposed method incorporates an efficient hybrid multiuser detection technique and consists of a two-stage procedure to achieve the performance of the ML detector with an acceptable level of computational complexity. The first stage performs interference cancellation by using the sorted QR decomposition (SQRD), and the second stage performs the genetic algorithm (GA). This two-stage procedure is thus called SQRD GA-MUD. The main advantage of the SQRD scheme is that it provides a "good initial setting knowledge" by eliminating IAI to improve the fitness of the population for GA. Simulation results obtained in this study demonstrate that SQRD GA-MUD achieves a gain of 6 dB to 15 dB compared to other well known detectors, with an acceptable number of iterations.