Convergence Properties and Algorithm Design for Iterative Receivers

Iterative receiver design has gained wide popularity in wireless communications. For algorithm design, the performance and the associated complexity are two major concerns. While both the performance and the complexity of an iterative algorithm are closely related to its convergence properties, in the existing studies, more often only the connections between convergence and performance are considered. The interplay between convergence and complexity, and therefore, e®ective complexity reduction, has not been fully appreciated. In this work, we consider typical receiver structures and develop e±cient algorithms based on their convergence properties.We start with iterative decoding for binary turbo codes, and show that by proper identification of decoding convergence after each iteration, the original optimal performance by full iterations can be maintained at a reduced complexity. A partial iteration algorithm ispresented. We compare the complexity, memory requirement, and implementation issues of the new algorithm with existing algorithms to show its advantages. An extension to serial concatenated codes is also provided. We also discuss other low complexity decoding strategies.We further extend the methodology to more general iterative receivers. As an example, we consider equalization for coded transmission over frequency selective channels. A partial iterative equalization and channel decoding scheme is introduced. We consider convolutional coded and turbo coded transmissions in details, and specify the corresponding receiver structures. The benefits from the proposed scheme are shown in various simulations. In the last part of the dissertation, we discuss iterative detection for multiple input multiple output (MIMO) transmissions. Speci¯cally, we consider sphere decoding based list detection algorithms for MIMO receivers. We propose a new algorithm that compared with existing algorithms, is simpler to implement but has better potential in performance. The convergence properties of this scheme is analyzed and compared with a known scheme. Finally, we point out several topics that are worth of further investigation.

Author

• Wu, Jinhong

Language

• en

Keyword

• Iterative Receiver
• Convergence
• Turbo Codes
• Iterative Equalization
• MIMO Detection
• Algorithm Design

Date created

• 2010

Type of Work

• Dissertation

License

• All rights reserved

Rights statement

• In Copyright

GW Unit

• Electrical Engineering

Degree

• Ph.D.

Advisor

• Doroslovacki, Milos I.
• Vojcic, Branimir R.

Committee Member(s)

• Subramaniam, Suresh
• Souryal, Michael R.
• Helgert, Hermann J

Persistent URL

•  https://scholarspace.library.gwu.edu/etd/jd472w616