Abstract
With a given set of measurements of instantaneous channel mixture of multiple sources, blind source separation (BSS) algorithms require only statistic of sources rather than training symbol to separate all source signals. Therefore, those blind algorithms are better spectrum efficient than those non-blind algorithms. Because there are specific assumptions for existing BSS algorithms, in this thesis, we make some arrangements and comparisons for some common BSS algorithms. Chi and Chen’s fast kurtosis maximization algorithm (FKMA), which is one of the BSS algorithms, by kurtosis maximization effectively extracts source signals using only spatial processing. By empirical studies we found that the smaller the normalized kurtosis magnitude of the extracted source signal, the worse the performance of the FKMA. Therefore, Chi et al. proposed turbo source separation algorithm (TSEA) also by kurtosis maximization which extracts one source signal through a spatial processing (for source extraction), and a temporal processing (for conversion of the extracted source into a filtered source with larger normalized kurtosis magnitude) cyclically. In this thesis, we analyse the performance of the TSEA and show some simulation results to support our analyses. For the extraction of all the unknown sources using either FKMA or TSEA, the widely used multistage successive cancellation (MSC) procedure has been an effective approach in spite of error propagation effects accumulated from stage to stage. Furthermore Chi et al. proposed two non-cancellation multistage (NCMS) BSS algorithms which extract all the unknown sources without any cancellation procedure. In this thesis, we analyse the performance of the two algorithms and simplify the mathematical equations. Chi and Chen’s FKMA has been successively applied to blind source separation, blind beamforming, and blind equalization. In this thesis, we apply the FKMA to a conventional cascade space-time receiver (CSTR), which involves a spatial processor to suppress co-channel interference (CCI) and a temporal processor to reduce inter-symbol interference (ISI). However, the performance of the receiver using FKMA is limited by the normalized kurtosis magnitude of the ISI distorted source signal. For this reason, we propose a novel receiver structure, referred to as turbo space-time receiver (TSTR). The proposed receiver is not sensitive to the normalized kurtosis magnitude of the ISI distorted source signal, so the proposed TSTR outperforms the CSTR. Finally, some simulation results are provided to support the performance of the proposed TSTR.