Mohammadreza Marandian Hagh

Iterative Receivers for Wireless Communication Systems"

Date: January 4, 2007

Abstract:

This dissertation is focused on low complexity implementation of iterative receivers for wireless communications. Increasing demand for high data rate services in limited bandwidths requires wireless systems to use more sophisticated techniques to utilize the channel capacity to provide more reliable services. Iterative receivers yield near capacity performance, but their implementation faces many challenges.

In the first part of this research, we address the implementation issues of turbo codes for the 3GPP wireless system. We introduce a low complexity new soft-input soft-output decoding algorithm as an alternative to the LogMAP algorithm. We also address the problem of excessive memory requirement in turbo codes by using the sliding window algorithms. We compare various iteration stopping algorithms to reduce the delay and power consumption of the turbo decoder.

Many wireless communication channels are subject to dispersion due to multipath propagation. Turbo equalizers are proposed to combat inter-symbol interference in dispersive channels. However, when channel memory length is large or a high order modulation is employed, turbo equalizers becomes impractical due to their excessive complexity.

In the second part of the thesis we investigate the problem of low complexity implementation of the iterative equalization techniques. We introduce a new iterative equalizer called the iterative decision feedback equalizer. This equalizer shows significant gain over non-iterative algorithms. The complexity of the iterative DFE is very low compared to the turbo equalizer. We study the performance of this equalizer using simulation and the extrinsic information transfer (EXIT) charts. We also show that for higher order modulation schemes, a new labeling technique can provide additional gain at higher signal to noise ratios compared to the standard Gray labeling.

Thesis Committee:
Professor John Proakis
Professor Masoud Salehi (Advisor)
Dr. Zoran Zvonar (Analog Devices)