Adaptive Reduced-Rank Minimum Symbol-Error-Rate Receive Processing for Large-Scale Multiple-Antenna Systems

摘要

In this work, we propose a novel adaptive reduced-rank receive processingstrategy based on joint preprocessing, decimation and filtering (JPDF) forlarge-scale multiple-antenna systems. In this scheme, a reduced-rank frameworkis employed for linear receive processing and multiuser interferencesuppression based on the minimization of the symbol-error-rate (SER) costfunction. We present a structure with multiple processing branches thatperforms a dimensionality reduction, where each branch contains a group ofjointly optimized preprocessing and decimation units, followed by a linearreceive filter. We then develop stochastic gradient (SG) algorithms to computethe parameters of the preprocessing and receive filters, along with alow-complexity decimation technique for both binary phase shift keying (BPSK)and $M$-ary quadrature amplitude modulation (QAM) symbols. In addition, anautomatic parameter selection scheme is proposed to further improve theconvergence performance of the proposed reduced-rank algorithms. Simulationresults are presented for time-varying wireless environments and show that theproposed JPDF minimum-SER receive processing strategy and algorithms achieve asuperior performance than existing methods with a reduced computationalcomplexity.