Successive Interference Cancellation for Uplink CDMA

Ph.D. Candidate
Department of Electrical Engineering
Stanford University, CA

Monday, April 15th, 11:00 AM, ENS 637

Abstract

Successive interference cancellation (SIC) is a promising technique for increasing the capacity of uplink CDMA systems due to its low complexity, allowance of strong error-correcting codes, and robustness in an asynchronous environment. Despite these attractive qualities and a decade of research, SIC has been plagued by several crucial problems that have prevented its adoption by industry. This talk presents effective solutions to these problems.

First, an optimum power control distribution is derived, that is specifically designed to accommodate estimation error. This power control formula is a general result for CDMA; previous results are special cases. It is shown that by using this distribution, even estimation error of up to 50% can be tolerated in using SIC to more than double the overall system capacity.

Second, a non-uniform received power distribution apparently implies greater power control complexity, as all users must be kept at or near their optimum power levels, which differ based on their decoding order. Nevertheless, a simple iterative technique for achieving the optimum distribution is shown to exist and converge.

Finally, a multipath channel seriously hinders previous SIC systems, suggesting the need for an integrated OFDM-front end. Analytical BER expressions for a novel Multicarrier-SIC (MC-SIC) system are derived in a frequency-selective fading channel, and simulation results confirmed the effectiveness of using narrowband carriers to combat multipath. There is virtually no degradation relative to an AWGN channel, and it is shown that the proposed MC-SIC system design with optimal power control increases capacity by an order of magnitude over current industry systems.

Biography

Jeffrey G. Andrews received the B.S. in Engineering with high distinction from Harvey Mudd College in Claremont, California in 1995, and the M.S. in Electrical Engineering from Stanford University in 1999. He is scheduled to complete his Ph.D. in Electrical Engineering at Stanford in June 2002.

From 1995 to 1997, he was an engineer at Qualcomm in San Diego, California, developing the Globalstar satellite communication system. He has served as a frequent consultant on wireless communications and signal processing to industry and the government, in particular to Microsoft, Ricoh, and NASA. His research interests are in wireless communication systems and communication theory.

A list of Telecommunications and Signal Processing Seminars is available at from the ECE department Web pages under "Seminars". The Web address for the Telecommunications and Signal Processing Seminars is http://signal.ece.utexas.edu/seminars