Joint estimation of array shape correction and noise source localization when the towed array turning around

The discrepancy between the estimated and actual shapes of the array during turns can adversely affect the precision of localizing near-field noise sources. This paper presents a solution to this challenge by employing Minimum Variance Distortionless Response (MVDR) focused beamforming to accurately identify the position of the near-field noise source based on an initial shape estimation. By utilizing the Taylor expansion, we establish a modified signal model that reflects the estimated array shape. A value function, along with its corresponding constraints, is then formulated to calculate the positional offsets of each array element accurately. The least squares method is employed to refine the array configuration, thereby determining the corrected positions of said elements. An iterative solution process ultimately yields both the noise source position and the adjusted array configuration. The simulation results conclusively demonstrate the effectiveness and accuracy of the proposed method, underscoring its potential for enhancing noise source localization in complex scenarios.