An efficient near-ML algorithm with SQRD for wireless MIMO communications in metro transportation systems

Abstract

A high-speed multiple input multiple output (NEMO) wireless communication system is proposed for metro transportation. In order to better mitigate the inter-antenna interference (LAI) which degrades the quality of the MIMO receiver, we propose an efficient hybrid multiuser detection (MUD) technique. This approach consists of a two-stage procedure to achieve the optimum multiuser detector (OMD) by an acceptable computational complexity. The first stage performs interference cancellation by using sorted QR decomposition (SQRD), and the second stage performs the genetic algorithm (GA). It has two significant advantages: 1) The SQRD scheme provides "good initial setting knowledge" to improve the fitness of the population for GA. 2) The effect of fitness calculation is obtained from the QR decomposition (QRD) of a MIMO channel. Simulation results demonstrate that the two-stage procedure obtains a gain of 3 dB to 25 dB than other well-known MUD schemes. The computational complexity of the two-stage procedure can be reduced by 30% with QRD than other fitness calculation scheme in GA-MUD.