| Since the early sixties, there has been considerable activity in the development of adaptive arrays for radar, sonar, communication, spectral estimation, etc. A key assumption in all previous work on adaptive beamforming is that the interfering signals are not coherent with the desired signal (i.e., one is not a scaled and delayed replica of the other). Coherence can completely destroy the performance of an adaptive array system. This is also the case for the recently developed high resolution methods, such as Maximum Entropy method, Minimum Variance method and eigenstructure method, for the localization of multiple sources. In practice coherent receiving environments often exist; for example, coherent interference can arise when multipath propagation is present, or when 'smart' jammers deliberately introduce coherent interference.; In this thesis we present a new processing scheme, applicable to the Maximum entropy, the Minimum Variance and the eigenstructure methods, that will retain their high resolution performance regardless of the coherence of the sources. The robustness to the coherence of the sources is achieved by spatial-smoothing; the array is divided into overlapping subarrays, and a smoothing operation is performed in the spatial-domain. Based on the spatial-smoothing idea, we also introduce a new adaptive beamformer able to work well even when the desired signal and the interference are coherent. The new adaptive processor uses both spatial and temporal sample data to overcome the degradation of performance in coherent receiving environments; it is also a cure for signal cancellation phenomena in adaptive arrays.; A statistical procedure called smoothed rank profile (SPR) test for determining the source coherent structure and the solvability of DOA estimation problem is proposed in the thesis.; In the thesis, an extension of the eigenstructure method is given for the case of incomplete information of the assumed signal subspace. A "signal implantation" technique is introduced to cope with the uncertainty model that accounts for phase and gain perturbations in sensors as well as for inaccurate knowledge of sensor positions in DOA estimation.; The last topic of the thesis is the overlapping echoes separation problem. An eigenstructure based method is suggested for estimating the number and arrival time of overlapping echoes with a priori known shape, from noisy observations received by a sensor. |
