Convergence study of iterative multi-user detection for turbo-coded CDMA

• Main Author: Hamidian, Behrooz
• Format: Others
• Published: 2006
• Online Access: http://spectrum.library.concordia.ca/8822/1/MR14260.pdf; Hamidian, Behrooz <http://spectrum.library.concordia.ca/view/creators/Hamidian=3ABehrooz=3A=3A.html> (2006) Convergence study of iterative multi-user detection for turbo-coded CDMA. Masters thesis, Concordia University.

In this thesis, we mainly focus on the convergence of IMUD turbo-coded systems. The main interest is the possibility that the system converges to the near-single-user performance and the second important concern is the speed of the successful convergence. The convergence thresholds of IMUD turbo-coded CDMA systems have been studied in [11] using conventional BER curves. However, BER curves do not have the ability to sufficiently describe the convergence behavior of iterative structures. Therefore in this thesis, we use the Variance Exchange Graph (VEG) as the methodology of analysis. VEG is a method that uses variance error of soft symbols as a measure to characterize the components in the IMUD structure. VEGs surpass BER curves for the purpose of convergence study by providing the possibility to analyze the system convergence from different angels. They also present a more detailed description of the iterative process comparing with BER curves. In this thesis, convergence thresholds and convergence speed of the proposed system are analyzed.