Research On Robust Adaptive Beamforming Algorithm Based On Array Steering Vector Estimation

Array signal processing has the advantages of strong interference suppression ability and high array gain.Much research results have been obtained now.Adaptive beamforming has good output performance in the ideal situation.However,due to the complexity of the underwater environment,there are disturbing errors,amplitude and phase errors,the existence of desired signals in the acquisition data,and the misalignment of steering vectors,which results in the decline of some existing algorithms in practical use.How to solve the problem of decreasing the performance of the algorithm becomes an important research hotspot in the field of current array signal processing.The robust adaptive beamforming algorithm with good robustness,fast convergence,low complexity is still the direction in which the array signal processing theory needs to be further developed.Aiming at the problems of array signal processing.Firstly,this dissertation do theoretical research and Matlab simulation analysis of CBF algorithm,MVDR algorithm and Diagonal Loading algorithm,and compare the output performance of the three algorithms under different conditions.CBF algorithm can't effectively suppress interference;MVDR has good performance and interference suppression capability under ideal circumstances.However,due to the sensitivity of MVDR to errors,the performance of MVDR is degraded when the steering vector is in error.The fixed loading of diagonal loading algorithm limits its usage range.In order to improve the robustness of the algorithm under the condition of array steering vector mismatch,Robust Capon beamforming(RCB)based on vector estimation is studied.The RCB algorithm can estimate the target accurately when the array steering vector mismatch,it has good robustness,and the performance of the algorithm is better than that of the MVDR algorithm.But,the performance of RCB algorithm is degraded in the covariance matrix of the incomplete rank.Then an improved RCB algorithm is proposed.The improved RCB algorithm directly decomposes the array receiving date matrix,avoids the inversion of the covariance matrix of the incomplete rank,and only need to calculate the left singular matrix,reduces the amount of computation.The improved RCB algorithm can satisfy the low shot condition and effectively estimate the azimuth of target when the covariance matrix is not in rank.Because the vector array receives more information than the scalar array,the vector array processing effect is better than the scalar array under the same condition,so the CBFalgorithm,MVDR algorithm and improved RCB algorithm are extended to vector array processing.The simulation analysis compares the output SINR size of vector and scalar algorithms and proves that the vector algorithm can achieve higher gain.Then,compared with the analysis direction spectrum,the vector algorithm has lower side lobe level,and the vector sensor array has the capability left-right resolution.Finally,the experimental data verify the validity of the algorithm.