Abstract
The kurtosis maximization criterion has been effectively used for blind spatial extraction of one source from an instantaneous mixture of multiple non-Gaussian sources, such as the kurtosis maximization algorithm proposed by Ding and Nguyen, and the fast kurtosis maximization algorithm (FKMA) proposed by Chi et al. By empirical studies we found that the smaller the normalized kurtosis magnitude of the extracted source signal, the worst the performance of these algorithms. In this thesis, with the assumption that each source is a non-Gaussian linear process, a novel blind source extraction algorithm, called turbo source extraction algorithm (TSEA), is proposed. The ideas of the TSEA are to exploit signal temporal properties for increasing the normalized kurtosis magnitude, and to apply spatial and temporal processing in a cyclic fashion to improve the signal extraction performance. The proposed TSEA not only outperforms the FKMA, but also shares the convergence and computation advantages enjoyed by the latter. This thesis also considers the extraction of multiple sources, also known as source separation, by incorporating the proposed TSEA into the widely used multistage successive cancellation (MSC) procedure. A problem with the MSC procedure is its susceptibility to error propagation accumulated at each stage. Therefore, we propose two noncancellation multistage (NCMS) algorithms, referred to as NCMS-FKMA and NCMS-TSEA, that are free from the error propagation effects. In this thesis, the FKMA is further applied to blind multiuser detection and blind space-time decoding (BSTD) for multicarrier code-division multiple access (MC-CDMA) systems. Assuming that all the users' spreading sequences are given \emph{a priori}, a blind multiuser detection algorithm (BMDA), which comprises FKMA and a user identification algorithm, for the uplink of a quasi-synchronous modified MC-CDMA system (with multiple receive antennas) and a BSTD algorithm, which comprises FKMA and blind maximum ratio combining (BMRC), for the down-link of a space-time coded MC-CDMA system (with multiple transmit and receive antennas used) are proposed. Moreover, the FKMA is applied to blind multiuser detection for asynchronous multi-rate direct sequence/code division multiple access (DS/CDMA) systems. The ideas are to properly formulate discrete-time multiple-input multiple-output (MIMO) signal models by converting real multi-rate users into single-rate virtual users, followed by the use of FKMA for extraction of virtual users' data sequences associated with the desired user, and recovery of the data sequence of the desired user from estimated virtual users' data sequences. Therefore, two multi-rate BMDAs (with either a single receive antenna or multiple receive antennas), which also enjoy the merits of super-exponential convergence rate and guaranteed convergence of the FKMA, are proposed in the thesis, one based on a convolutional MIMO signal model and the other based on an instantaneous MIMO signal model. For blind co-channel interference and intersymbol interference (ISI) reduction in cellular wireless communications, the FKMA is applied to the design of the conventional cascade space-time receiver (CSTR) in this thesis. However, the receiver performance is limited as the normalized kurtosis magnitude of the ISI-distorted signal of interest is small. Then a blind turbo space-time receiver (TSTR) is further proposed that applies spatial and temporal processing using FKMA in a cyclic fashion to estimate the desired data sequence. The performance of the proposed blind TSTR is insensitive to the value of the normalized kurtosis magnitude of the ISI-distorted signal of interest, and therefore is superior to that of the blind CSTR. This thesis also considers blind beamforming of multiuser orthogonal frequency division multiplexing (OFDM) systems. Assuming that the channel is static within one OFDM block, a blind beamforming algorithm by kurtosis maximization based on subcarrier averaging over one OFDM block is proposed, which basically comprises source extraction, time delay estimation and compensation, classification, and BMRC. The designed beamformer is exactly the same for all the subcarriers, effectively utilizes multipath diversity for performance gain, and is robust against the effects of correlated sources. Finally, some simulation results are presented to demonstrate the effectiveness of the proposed blind source separation algorithms, BMDAs, BSTD algorithm, blind CSTR and TSTR, and blind beamforming algorithm.