Iterative Methods for Multiuser Detection: Impulsive Noise, Coded Channels, and Multiple-Antenna Sytems

Princeton University, Princeton NJ

Investigators

Abstract

Multiple-access communication channels arise in a great variety of applications, among which are some of the most rapidly growing areas of technology development today: cellular telephony, wireless "last mile'' systems, indoor wireless LAN's, nomadic computing, high-speed digital subscriber lines, etc. New performance requirements are constantly being imposed on such systems in order to provide high-quality, multimedia services to an exponentiating user population. A major approach to meeting these requirements is through the use of advance signal processing techniques to mitigate the manifold physical impairments arising in such channels. This research addresses a number of key issues in this field. Multiple-access channels arise in the above-noted systems as the result of several physical conditions, including non-orthogonal multiplexing in CDMA systems, cross-talk in xDSL systems, multipath in TDMA systems, and co-channel interference in most types of cellular systems. Multiuser detection, developed over the past fifteen years, has provided a very useful theoretical basis for the study of optimal signal processing problems in multiple-access channels. This research is concerned with problems in which multiuser detection must be combined with other signal processing functions, such as beamforming, channel decoding, impulse-noise mitigation, etc. The basic approach is to apply algorithms that iterate between signal-processing functions. In particular, this research examines the general structural and performance properties of such algorithms, their implementation, and their application in three specific technologies: indoor wireless networks, digital subscriber lines, and underwater acoustic modems.