Research On Robust Beamforming Algorithm Based On Sparse Reconstruction And DSP Parallel Implementation

As an important issue of array signal processing,adaptive beamforming has been widely applied in radar,sonar,satellite navigation,wireless communication,medical image,and speech recognition,etc.However,in practical applications,such factors,like channel amplitude and phase perturbations and antenna displacement,which make adaptive beamformer constrain target steering vector inaccurately,will result in array output performance degradation due to achievement on weight vector by target-involved sample covariance matrix.To solve the aforesaid problems,this thesis studies robust adaptive beamforming in detail.The main work of this thesis is as follows:Staring from the principle of Capon beamformer,several state-of-art robust beamforming algorithms,which belong to the family of loading methods,the family of subspace methods,the family of constraint optimization methods,and the family of sparse reconstruction methods,are studied in this dissertation.At first,the basic ideals and enforcement steps of the above robust beamforming algorithms are specified;and then,the advantages and shortcomings of the aforesaid robust beamformers are analyzed;finally,the performances of the stated robust adaptive beamforming approaches are verified through numerous experiments.The experiment results show that the existing robust algorithms cannot constrain target steering vector precisely,in the situation of direction of arrival error and array structure mismatch,which leads to the output performance severely declining.Focusing on the problem of output signal-to-interference-plus-noise ratio losses of both the family of sparse reconstruction approaches and the family of constraint optimization approaches,this thesis proposed the steering vector and covariance matrix joint iterative estimation for robust beamforming algorithm.The initial value of target steering vector is obtained by spectrum reconstruction,and then the interference-plus-noise covariance matrix is initializd through eliminating the target from the sample covariance matrix.To proceed,on the basis of the steering vector constrain optimization model,the optimal target steering vector and interference-plus-noise covariance matrix are joint-iteratively solved.Experimental results show that the joint iterative estimation algorithm can improve the output performance of beamformer in the situation of array structure error.Aiming at the problem that the joint iterative estimation algorithm cannot accurately estimate the interference-plus-noise covariance matrix when the mismatch of the target steering vector is large,this thesis proposed the interference-plus-noise covariance matrix reconstruction via target robustly blocking for beamforming and the robust beamforming via alternating iteratively estimating the steering vector and interference-plus-noise covariance matrix.The former algorithm pre-processes the training data by constructing the direction of arrival extension-based target blocking matrix,and then achieves the interference-plus-noise covariance matrix reconstruction via matrix transformations.The later method uses spectrum reconstruction to get the initial value of target steering vector,and then completes the initialization of the interference-plus-noise covariance matrix by constructing target blocking matrix.To proceed,on the basis of the steering vector constraint optimization model,the optimal target steering vector and interference-plusnoise covariance matrix are alternate-iteratively solved.Simulation results illustrate that the covariance matrix via target robustly blocking for beamforming performs little worse than the joint iterative estimation algorithm in spite of the preferable computation complexity.Moreover,the alternate iterative estimation beamforming algorithm has robustness improvement compared to the joint iterative estimation algorithm while it is not applicable for weak target scenario.On the basis of the TMS320C6678 DSP hardware platform,this thesis also devises the implement scheme of the proposed robust beamforming algorithm.Upon the given scheme,the entire procedure and function modularizations of the proposed robust algorithm are thoroughly illustrated,which follows the hardware experiments to evaluate effectiveness of the designed scheme.Simulation results demonstrate that the robust beamforming scheme can achieve gain on the target and suppress the interference simultaneously.