Research Of Robust Adaptive Beamforming Based On Non-Uniform Sparse Array

Adaptive beamforming is a core technique in array signal processing field.In order to suppress interferences and receive signal of interest correctly,Adaptive beamformer changes the weight vector of the antenna sensor adaptively by analyzing the received signal.Recently,researchers aim to improve the robustness of adaptive beamformer,because that adaptive beamformer is known to be sensitive to model mismatch in practical environment.This thesis focuses on robust adaptive beamforming based on non-uniform sparse array,including the optimization of RAB algorithm itself and its application to non-uniform sparse array such as nested array and coprime array.At present,sparse arrays are widely used in array signal processing,but most research focus on DOA estimation.The research on adaptive beamforming is insufficient.Direct application of traditional RAB algorithm to sparse arrays will result in reduced output performance.In this thesis,the RAB algorithm based on non-uniform sparse array is studied in detail,and the problems encountered when applying RAB algorithm to non-uniform sparse array are solved as far as possible.The main work and contributions of this thesis can be summarized as follows:This thesis firstly focuses on uniform linear array and the effectiveness of the current three INCM reconstruction method is analyzed.The influence of interference power estimation on INCM reconstruction is also revealed.It is pointed out that overestimated interference power has little effect on the performance of beamformer.Then,based on the analysis results,a RAB algorithm based on fast power estimation and INCM reconstruction is proposed.This algorithm can effectively reduce the complexity of power estimation and guarantee the performance of beamformer.Next,in the study of nested array,the convergence property of the spatially smoothed matrix is analyzed.It is pointed out that the inverse matrix of the spatially smoothed matrix constructed directly from the sample covariance matrix converges slowly,which results in the performance degradation of the adaptive beamformer.A nested array RAB algorithm based on interference spatially smoothed matrix is proposed.The algorithm not only has a fast convergence rate,but also removes the desired signal components in the sample data,thus avoiding the signal self-nulling effect under high SNR situation.At last,considering the fact that the DOFs of the difference co-array is not completely used in existing coprime array RAB algorithms,a RAB algorithm of coprime array based on matrix completion is proposed.In this algorithm,matrix completion technique is used to interpolate the difference co-array of the coprime array,which improves the available DOFs of the co-array and improves the output performance of the adaptive beamformer indirectly.