Blind adaptive array techniques for mobile satellite communications

This thesis is concerned with the investigation of blind adaptive array techniques for mobile satellite communications. The algorithms developed in this thesis are block iterative in nature and exploit temporal properties of certain communication signals. The term ‘blind’ in this context is meant to indicate that the use of training sequences or knowledge of the direction of arrival for the signal of interest is not required. As part of the investigation, the algorithms are implemented over realistic simulated satellite links in the presence of modeling errors and variations in the array subsystem.; Chapters 2 and 3 are dedicated to the characterization of distortions caused by the propagation environment and array operating environment, respectively. Chapter 2 reviews the basic propagation models reported in the literature for wireless communications. Chapter 3 focuses on the effects of deterministic or systematic errors introduced by the array subsystem. In particular, the effects of nonlinearity and channel shaping on the performances of previously reported blind adaptive array algorithms are noted.; Chapter 4 marks the beginning of the innovative array processing research presented in the dissertation. In this chapter, a detailed convergence analysis of blind FA beamformers is presented. To provide insight into the convergence behavior of these multidimensional algorithms, the notion of principal angles is introduced. Principal angles are a generalization of the concepts of angles between planes extended to sub-spaces.; Chapter 5 introduces a novel space-time algorithm that exploits the bursty nature of the wireless packet data services for satellite links. More specifically, the algorithm exploits the capability of algebraic codes to indicate packet error rates for the purpose of data partitioning and signal separation via least squares estimation combined with signal projection techniques.; Finally, a null steering beamspace algorithm is proposed for the multiuser detection of orthogonal frequency division multiplexing (OFDM) signals. The algorithm uses a linear predictive beamformer to minimize the output power in the orthogonal null space of the signal of interest (SOI). A robust version of the algorithm is introduced to prevent signal cancellation when synchronization errors are present, and a partially adaptive version of the algorithm is also proposed to reduce the computational complexity of the algorithm. (Abstract shortened by UMI.)...