Research Of Robust Beamforming And Direction Estimation With Convex Optimization In Uncertain Ocean Environment

In this thesis,the robust adaptive beamforming and signal direction estimation for passive sonar systems in complex uncertain ocean environments are studied.The main work is summarized as follows:1)A robust adaptive beamforming algorithm based on spatial power spectrum match-ing is proposed.First,it constructs a convex optimization problem based on spatial power spectrum matching,and realizes the reconstruction of the desired signal covariance matrix and the interference plus noise covariance matrix.Second,it applies the reconstruction matrices to the generalized rank minimum variance distortionless response method.Fi-nally,the robust adaptive beamforming algorithm for the generalized rank source model is obtained.A robust adaptive beamforming method based on spatial power spectrum matching can still achieve robust adaptive beamforming with less a priori information,fewer snapshots,and lower signal to noise ratio2)A robust adaptive beamforming algorithm based on subspace optimization is pro-posed.It uses subspace processing and convex optimization techniques to achieve ac-curate estimation of the steering vector and reconstruction of the interference plus noise covariance matrix.Both simulation and experimental data show that the robust adaptive beamforming algorithm based on subspace optimization has good robustness in terms of direction of arrival error,random wavefront perturbation,array element position error,coherent local scattering and incoherent local scattering3)A robust adaptive beamforming algorithm based on semi-positive definite relax-ation is proposed.It expresses the estimation of the steering vector as a non-convex quadratic constrained quadratic programming problem,and obtains the approximate opti-mal solution of the steering vector by the semi-positive definite relaxation and the random-ization processing.The robust adaptive beamforming algorithm based on semi-positive definite relaxation achieves the best SINR results under various uncertainties4)A direction-of-arrival estimation algorithm based on steering vector matching search is proposed.It searches for the matching steering vector so that the minimum value of the beam amplitude response of the MVDR beamformer is maximized in the ob-servation interval.The direction-of-arrival estimation algorithm based on steering vector matching search does not require the information of signal subspace or noise subspace,and achieves significantly higher estimation accuracy than the traditional direction-of-arrival estimation algorithms with low SNRs,less snapshots,and strong interferences5)A direction-of-arrival estimation algorithm based on matrix sparse low rank de-composition is proposed.The signal covariance matrix is low-rank but non-sparse,and the noise covariance matrix is sparse but not low-rank.The convex optimization problem is thus constructed to achieve sparse low-rank decomposition of the sample covariance matrix.The direction-of-arrival estimation algorithm based on matrix sparse low rank decomposition achieves higher resolution than traditional methods.