Turbo Multiuser Detection: Analysis, Optimization and Applications

Columbia University, New York NY

Investigators

Abstract

Wang, Xiaodong Columbia University Objectives and Significance: The turbo principle [27 ],namely,the strategy that exploits the iterated exchange of soft information between di .erent blocks in a communication receiver,has grown to be a powerful tool in attacking a diverse set of problems in communications.In particular,the turbo multiuser detection (MUD)paradigm [59 ]has attracted signi .cant recent attention as an e .ective technique for joint decoding in multiuser communication systems.However to date the performance of various turbo MUD schemes is largely demonstrated via simulations,and the impact of turbo MUD on the network performance (e.g., throughput and delay)has not been explored systematically.It also remains unclear at this time how the upper-layer functionalities (such as ARQ protocol,power control,and admission control)should be designed to exploit the physical-layer turbo MUD.The objective of the proposed project is to develop analytical tools to assess and optimize the performance of various turbo MUD methods,and to apply these esults to c oss-layer design in wireless networks. Recently a technique known as density evolution [66,67 ]has been developed in the coding community that permits analysis of iterative decoding algorithm for in .nite-length codewords.In this proposal,under the density evolution framework,we outline a research plan aimed at understanding of the design and analysis of turbo multiuser detection for a variety of channel conditions and receiver structures,and the design of some upper-layer functionalities which takes into account the physical layer turbo MUD.While in this study we focus on multiuser CDMA systems,the general principle is applicable to several systems,such as space-time coding systems,OFDM systems,and intersymbol interference systems.