Low-Complexity Detection Techniques for MIMO and Cooperative Networks

Abstract

This thesis investigates detection technologies for multiple-input multiple-output (MIMO) systems and cooperative communications. The exploitation of detection methods and relay strategies to achieve near-optimal performance (measured by symbol error rate or minimum mean squared error) and reductions of the running time of detection methods are the main focus of this thesis. A signal-to-noise ratio (SNR)-dependent radius control sphere detector (SD) algorithm, a general framework of statistical pruning SD, and an improved K-best SD are proposed. These SD algorithms reduce the detection running time in terms of the average number of visited nodes, with negligible performance loss compared with that of optimal maximum likelihood (ML) detection. In order to optimize the MIMO relay performance, an estimate-and-forward (EF) relay strategy is also developed. The emerging trend towards large MIMO and cooperative communication systems makes the development of low running time strategies with near-optimal performance more important. Thus, an EF list generated by SD is also proposed to reduce the number of computational operations for the EF scheme in large MIMO relay networks; this method is called list EF. Overall, the research findings should help to reduce the running time and improve the reliability of detection algorithms, to achieve a desirable trade-off between running time and performance, and to provide efficiently-implementable MIMO and cooperative detection algorithms.