Research On Robust Capon Beamforming Algorithms

Adaptive beamforming refers to enhancing a signal from a specific direction of arrival based on received signal.This process is similar to forming a "beam".Due to high signal interference to noise rate,it has been widely used in medical imaging,sonar,radar,astronomy and communications.The mismatch of data measured by sensors often causes drastic performance degradation in practice.Therefore,the research on robust adaptive beamforming algorithm has always been a hot issue.The problem that needs to be solved in the robust adaptive beamforming algorithm is the performance degradation caused by covariance matrix mismatch and signal steering vector mismatch.The mismatch of the covariance matrix and the mismatch of the signal steering vector cannot be eliminated and such mismatch can be caused by environmental nonstationarities,look direction errors,array calibration errors,signal wavefront distortions,array geometry distortions and local scattering,etc.The research of enhancing robustness is mainly aimed at these two kinds of mismatch.In order to improve the robustness of adaptive beamforming,the research contents of this paper are as follows:(1)Eleven robust adaptive beamforming algorithms are studied and divided into three categories:robust adaptive beamforming algorithms based on loading matrix,robust adaptive beamforming based on eigen-subspace and robust adaptive algorithms based on steering vector uncertainty set.Six of them are simulated by math software.(2)The relationship between weight vector and robustness in beamforming algorithm is studied,The concept of design algorithms is derived that the optimal weight vector of the strong robust adaptive beamforming algorithm must be close to the estimated signal steering vector to obtain robustness,but it must not be completely equal to the estimated signal steering vector to avoid losing the adaptivity.Based on these two research,two robust adaptive beamforming algorithms are proposed in this thesis:(1)A robust adaptive beamforming with null-pattern constraints is proposed.Linear constraints and norm constraint of weight vector are adopted to suppress interference signal and against signal steering vector uncertainty respectively.It is proved through mathematical transformation that the norm of the weight vector is amonotonically decreasing function on Lagrangian multiplier.Simulations in which three algorithms are involved for comparison show that the proposed algorithm has better performance.(2)Another algorithm called negative diagonal loading robust adaptive beamforming is proposed.It is based on standard Capon beamforming and the objective of optimization problem is to maximize the output power with the constraint that weight vector close to the signal steering vector to obtain strong robustness.It's demonstrated by simulations that the proposed algorithm outperforms five algorithms in scenes with varying snapshot numbers and scenes with large deviation of signal steering vectors.