Joint Array Shape And Bearing Estimation For Towed Linear Array Sonar During Maneuvering

The direction of arrival(DOA)estimation technique is widely utilized in radar,sonar,and com-munication for target localization and tracking.In underwater acoustic detection,high-resolution DOA estimation techniques are required to image the target and improve the recognition rate.This thesis focuses on the target localization problem for towed line array during maneuvering,aiming to improve the deconvolved conventional beamforming(DCBF)and sparse Bayesian learning(SBL)methods for high-resolution DOA estimation.The conventional DCBF method achieves high-resolution directivity by deconvolution in the spatial domain and eliminating the influence of array shape uncertainty.To improve the DOA es-timation performance at low signal-to-noise ratios,the DCBF algorithm for the signal subspace is studied.Then,the DCBF-based dual-subarray fusion method for DOA estimation and the subarray decomposition method for array shape estimation are proposed.The former obtains an accurate spatial spectrum and resolves the left-right ambiguity of the target bearing by dividing the array into two subarrays to reduce the location errors.The latter estimates the array position by the dif-ference in DOA estimation of each neighboring subarray,which solves the problem of performance degradation for towed line arrays due to array mismatch during maneuvering.The SBL algorithm is utilized for multi-target DOA estimation research under the conditions of strong interference and low signal-to-noise ratio,which transforms the DOA estimation problem into a sparse reconstruction problem,and obtains the high-resolution spatial spectrum by combin-ing a priori information with the observations.The method of using sparse line arrays instead of uniform line arrays is proposed for the DOA estimation of long towed line arrays to reduce the computational burden due to multiple array elements and snapshots.Meanwhile,the array shape and the direction can be jointly estimated from the acoustic data,which effectively improves the detection performance of weak targets for towed line arrays under maneuvering conditions.The simulation results verify that the DCBF method and the SBL algorithm are both robust and sensitive,suitable for DOA estimation with mismatched array shapes and a limited number of snapshots.The Zhoushan sea trial experiment verifies the practicality of the DCBF method and the SBL algorithm for underwater target detection.The processing of MAPEX2000 sea trial experimental data verifies that the SBL algorithm is qualified for towed line array sonar to overcome the interference and estimate the DOAs of weak targets with high precision during maneuvering.