Robust Array Signal Processing Techniques And Their Applications

Under ideal conditions, adaptive array signal processing methods can get excellent performance. In practical operating circumstances, the performance of adaptive array signal processing methods degrade extremely due to existing errors. Furthermore, adaptive array signal processing methods usually require a lot of computations. These prevent greatly their applications in practical systems. Therefore the study of robust and fast algorithms for adaptive array signal processing is becoming more urgent and important. The focus in this dissertation are on discussion of some robust and fast algorithms for adaptive array signal processing. The main content can be outlined as follows: We propose a modified eigenspace-based adaptive beamforming(MEBAB) algorithm and a fast eigenspace-based adaptive beamforming(FEBAB) algorithm to overcome the drawbacks of the eigenspace-based adaptive beamforming(EBAB) algorithm. The MEBAB algorithm first calibrate the constraint steering vector, and then use the calibrated steering vector for beamforming. It is more robust to beam pointing error than the EBAB algorithm. The FEBAB algorithm can directly obtain the signal plus interference subspace from the elements of the array covariance matrix, and so it has a low computation complexity in comparison with the EBAB algorithm. We discuss the generalized eigenspace-based beamforrning(GEIB) algorithm and propose the eigenspace-based linearly constrained minimum variance beamforming(ELCMVB) algorithm. To preserve the constraints, the GEIB algorithm has to compute a modified signal plus interference subspace. This may result in numerical instability and the performance of the GEIB algorithm is affected by the constraints. The ELCMVB algorithm can overcome these problems and performs better than the GEIB algorithm. Computer simulation results demonstrate the merits of the ELCMVB algorithm. Because coherent signals may often present in practice. We present a new method of optimum beamforming in the presence of coherent signals. The method first gets the composite steering vector of the desired signal and the coherent signals, and then projects this vector onto orthogonal complementary space of the uncorrelated interference subspace as the weight vector. The theoretical analysis indicates that the proposed method is nearly identical to the optimum method. The proposed method can be applied to an array of arbitrary Iv ABSTRACT geometry and is robust to the estimation errors of the source directions. Based on the eigenanalysis interference cancellation method, two efficient interference cancellation methods are presented. The first method gets the modified interference subspace directly from the elements of the array covariance difference matrix and obtains the adaptive weight vector by utilizing this subspace. Compared to the eigenanalysis interference cancellation method, the first method reduces the computation complexity and is robust to the array complex gain error. Based on the first method, the second method first calibrates the constraint steering vector, and then does interference cancellation. The second method is not only robust to the array complex gain error, but also to beam pointing error. To reduce the computation complexity of the orthogonal projection algorithm for adaptive beamforming, we propose a fast orthogonal projection algorithm. The proposed algorithm can obtain the modified interfe...