Multiple-Symbol Detection Aided Differential Spatial Division Multiple Access

Abstract

This paper presents a multiple-symbol differential spatial division multiple access (MS-DSDMA) system conceived for low-complexity and high-bandwidth-efficiency applications operating in time-varying fading channels, where no channel estimation is required. A low-complexity adaptive multiple-symbol differential interference suppression (MS-DIS) scheme is proposed, which is based on the maximum signal-to-interference-plus-noise ratio (MSINR) criterion and facilitates the implementation of the powerful multiple-symbol differential sphere detection (MSDSD). Then, a practical three-stage turbo DIS receiver design framework is proposed for the MS-DSDMA system, which is constituted by concatenating the adaptive DIS filter bank, the MSDSD and the channel decoder. Both the EXtrinsic Information Transfer (EXIT) chart analysis and the Monte-Carlo-based simulation results show that the proposed three-stage turbo DIS scheme is capable of achieving a substantially enhanced performance in comparison to the conventional linear minimum mean-squared error (LMMSE) based adaptive receiver. Furthermore, for the sake of significantly reducing the iterative detection complexity, two complexity reduction techniques are devised, namely the a priori-LLR-threshold (ALT) and the adaptive-window-duration (AWD) aided MSDSD schemes.