Research On Robust Adaptive Beamforming

As one of the most important research directions in array signal processing,beamforming technology has been realized in practical engineering applications based on the progress of digital processor technology.Beamforming technology has been widely used in sonar,mobile communication,radar,radio astronomy and geological detection.The beamformer performs a complex weighted summing operation on the sampling sequences of each channel to receive the desired signal and suppress the interference signal.Capon beamformer is one of the most widely used adaptive beamformers.Compared to the conventional beamformer,Capon beamformer has higher resolution,array gain and better interference suppression capability.However,the existence of non-ideal factors such as array structure error,channel disaccord,and multipath propagation in real systems will lead to performance degradation.When the desired signal components are presented in the sampling sequences,the Capon beamformer will suffer more severe performance degradation.The main research contents of this thesis are listed as follows:1)Robustness analysis of beamforming algorithms.First,we analyze the signal model and design principle of Capon beamformer.Then,the theoretical basis of robustness problems caused by steering vector(SV)mismatch and covariance matrix mismatch is analyzed in detail.By analyzing the design ideas and principles of various commonly used robust beamforming algorithms in detail,we summarize the disadvantages of these algorithms.2)Research on robust adaptive beamforming based on Covariance matrix reconstruction with iterative mismatch approximation.Conventional robust beamforming algorithms only focus on improving the robustness problems attributed to the desired signal SV and covariance matrix constructed by finite sampling sequences mismatches,which makes the Capon beamformer unable to cope with performance degradation caused by the desired signal "self-nulling".To solve this problem,we propose a new robust adaptive beamforming algorithm based on interference plus noise covariance matrix(INCM)reconstructed by Iterative Mismatch Approximation(IMA)method.In this algorithm,we employ the Capon spatial power spectrum to estimate each direction of arrival(DOA)of interference signals and corresponding nominal SVs at first.Then,the INCM is reconstructed from the interference signal SVs and the power estimated by the IMA method.Finally,the IMA method is also used to estimate the desired signal SV and calculate the weight vector.To correct the SV mismatch using IMA method,we first establish the model of signal SV mismatch.Then,we iteratively estimate each SV element mismatch value in a pre-defined gain-and-phase mismatch region based on a maximum power criterion.Finally,we correct the nominal desired SV by using the estimated mismatch vector.The proposed algorithm is easy to implement without optimization toolbox.Besides,this algorithm is robust against arbitrary SV mismatch and has low computational complexity.3)Research on low complexity robust adaptive beamforming Based on INCM Reconstruction via Subspace Projection.Most conventional INCM reconstruction based robust adaptive beamforming algorithms reconstruct the INCM via estimating interference SVs by Capon spatial power spectrum or directly integrating Capon spatial power spectrum on the angular region in which all interferences lie in.This makes the computational complexity very high.Therefore,we propose a new low complexity robust adaptive beamforming algorithm based on INCM reconstruction via subspace projection.In this algorithm,we first construct a block matrix.Then,the block matrix is used to calculate the projection matrix for "filtering" the desired signal components from the signal subspace,and the INCM is subsequently reconstructed.Finally,the IMA method is used to estimate the desired signal SV and calculate the weight vector.The blocking matrix is constructed based on the stopband constrained passband minimum mean square error(MSE)criterion and the minimum MSE criterion.The projection matrix calculated from the blocking matrix removes the desired signal components while remaining the interference components.Since the construction of the blocking matrix is independent of sampling sequences,it means that calculating the blocking matrix can be seen as a preprocessing operation without increasing the complexity of the proposed algorithm.The algorithm has low complexity and good performance under high SNRs conditions.4)Research on robust adaptive beamforming Based on INCM reconstruction via Oblique Projection.The power leakage problem caused by the presence of noise and SV mismatches,as well as the estimation accuracy reduction problem caused by coherent source,will lead to a decrease in the accuracy of the INCM reconstructed based on the Capon spatial power spectrum.In addition,the computational complexity of Capon spatial power spectrum estimation is high.Therefore,we propose a new robust beamforming algorithm for INCM reconstruction based on oblique projection.In this algorithm,we employ the Minimum Norm of Subspace Projection(MNSP)method to estimate the desired signal SV and calculate the oblique projection operator in conjunction with the signal subspace first.Then,we employ the oblique projection operator to correct all nominal interference SVs estimated by the maximum entropy power spectrum.Finally,we reconstruct the INCM and calculate the weight vector.Based on the orthogonality between the noise subspace and the actual signal SV,the MNSP method estimates the desired signal SV by minimizing the norm of the SV projected on the noise subspace.The oblique projection operator is calculated from the estimated desired signal SV and signal subspace.The proposed algorithm has good performance under high SNR conditions.