Reduced-complexity iterative multiuser detection

• Main Author: Qin, Zhiliang
• Other Authors: Teh, Kah Chan
• Format: Theses and Dissertations
• Published: 2008
• Subjects: DRNTU::Engineering::Electrical and electronic engineering::Computer hardware, software and systems
• Online Access: http://hdl.handle.net/10356/3155
• Institution: Nanyang Technological University

In this thesis, we consider iterative multiuser detection for coded code-division multipleaccess (CDMA) systems. In contrast to most previous work that considers convolutionally coded systems, we propose iterative receivers for turbo coded and serially concatenated convolutionally coded (SCCC) systems. The proposed receiver consists of a full-complexity a posteriori probability (APP) multiuser detector and a bank of iterative channel decoders, which are tightly combined at each iteration. Simulation results show that near-single-user performance can be achieved at the expense of exponentially growing computational complexity with the number of users. To reduce the computational complexity of multiuser detection, we then focus on developing reduced search soft-input soft-output (SISO) multiuser detectors for synchronous coded systems. Based on the snapshot property of synchronous CDMA channels, the proposed multiuser detector limits APP multiuser detection to certain neighbors of tentative hard estimates fed back from channel decoders. As a further simplification of reduced-search multiuser detection, an iterative interference cancellation (IC) scheme is also investigated. Extensive simulation results are presented and semi-analysis is also conducted. It is shown that the proposed multiuser receiver can provide reasonable and adjustable computational complexity. Following that, we develop reduced-state SISO multiuser detectors over an asynchronous coded CDMA channel based on a channel trellis simplification approach. Tentative soft/hard information is used to select survivor states in the channel trellis. Compared with the full-complexity iterative receiver or single-user performance, the proposed receiver has small performance degradation and can provide an excellent performance/complexity tradeoff.