Research On Array Signal Processing And Its Related Technology

Array signal processing and its application has been one of the focuses in signal processing field. By sampling and processing signal both in time domain and in spatial domain, the information of interest contained in the signal can be exploited fully, thereby interference can be suppressed more effectively, and so the capacity of system can be improved by introducing array signal processing algorithms. In this dissertation, the model and theory of array signal processing are studied; DOA estimation and beamforming technology and their correlated technology were researched deeply, which are the two main aspects in array signal processing field.A geometrically based single bounce elliptical (GBSBE) multiple channel model is analyzed in this dissertation, based on probability distribution theory of stochastic variable function, a simulative method of this channel model is designed.MUSIC algorithm has a great disadvantage that it has large amount of calculation, a algorithm is proposed, which applies to uniform linear array, and this algorithm can decrease the amount of calculation of DOA estimation.In communication system, some artificial signals have a characteristic of cyclostationarity, the noise and interference suppression property of cyclostationarity statistic is studied. The research denotes that cyclostationarity statistics can suppress the signals which have different cyclic frequency.A DOA estimation algorithm based on this cyclic auto-relation theory of array signal is proposed. Simulations show that this algorithm has good performances in DOA estimation. The effect of mutual coupling on spectrum correlation subspace fit (SC-SSF) is analyzed, and then a compensating method is presented in this dissertation.Coherent signal is common in radio communication system, the problem of coherent signal DOA estimation is studied and compare the property of forward spatial smoothing with that of for-backward spatial smoothing.On the other aspect of array signal processing--adaptive beamforming, main researches are as follows.Because the Linearly Constrained Minimized Variance (LCMV) beamformer is sensitive to the error of DOA point a weighted LCMV beamformer is studied, which weights the weight vector by means of the outputting power ofbeamformers based on different DOA. Simulations indicate that this algorithm increases the outputting SINR of beamformer.In the case of little sampling snapshot, LCMV beamformer have shortcomings that sidelobe levels are too high and the weight vector jitters. To overcome the shortcoming, a method is presented which integrates the quiescent pattern constraint and adaptive algorithm. This method not only stabilizes the sidelobe levels but also lets the antenna array possesses adaptive ability.In the case of moving interferences or moving antenna array, the nulling of the antenna array cannot point to the DOA of interference perfectly, which causes the property of system spoiling. In this dissertation an algorithm that broadens the nulling is presented. This algorithm increases the robustness of antenna array system.Contrast and analyze the performances of minimizing mean square error (MMSE) beamforming algorithm and LCMV beamforming algorithm in the condition of coherent signals. Analysis denotes that the MMSE beamformer has a good performance in outputting SINR, but the outputting signal of LCMV beamformer will be submerged in noise and interferences. And then the performances of spatial smoothing algorithm to decorrelate are researched. Aimed to the case of little angle scatter, the manifold of array is extended to the sum the original array manifold and its weighted derivative, this kind of array manifold increases the freedom degree of adaptive algorithm, so the beamformer based on the improved manifold has the better outputting property than the beamformer based on conventional manifold.