Near-field High-resolution Localization And Identification Method Of Underwater Noise Sources Based On Vector Sensor Array

Acoustic stealth is one of the basic tactical indexes for the submarine. Through studying the localization and identification of the noise sources, we can better guide the noise-reduction and vibration-isolation. Along with the reduction of the noise level, the methods to evaluate the contribution of the noise sources need improvement. Because the noise sources may be in the near field of the array, the research on the localization and identification of the noise sources in the near field is very significant.Based on the array theory in the near field, this paper analyzes the factors which effects the application of the vector sensor array, furthermore, introduces the principle of the conventional focused beamforming method, and deduces the calculation formula for the focused spatial resolution.Because the resolution of the conventional focused beamforming is low, the vector sensor array MVDR high-resolution focused beamforming method (VMVDRFB) is proposed. The simulation results show that VMVDRFB has higher localization precision, stronger anti-interference capability, better robustness, higher spatial resolution, better spatial anti-aliasing performance, as well as stronger capability to distinguish the starboard and the larboard. So VMVDRFB is the successful integration of the spatial spectrum algorithm and the vector sensor array technique, and can effectively improve the performance of the noise source localization and identification. Meanwhile, a new method based on the near field interpolated array technique is studied. This method can distinguish the near field coherent sources without the multi-dimensional searches, and can reflect the relative contribution of each coherent source. Moreover, this method is simple and efficient.The vector sensor array broadband high-resolution focused beamforming algorithms are studied, including the sub-bands decomposition MVDR focused beamforming algorithm, which is applicable to the non-coherent broadband signals, as well as including the CSS-MVDR and the STMV focused beamforming algorithms, which are both applicable to the coherent broadband signals.A new near field source parameter estimation method based on the combination array is proposed, which can provide the reliable prior information for the focused algorithm. The form of the array is a uniform vertical pressure sensor array, in whose center there is one three-dimensional vector hydrophone. By this combination array, the three-dimensional parameters of the near-field sources including the azimuth angle, the elevation angle and the distance can be achieved. This method only uses the second-order statistics information, and transforms the three-dimensional parameter estimation problem into several one-dimensional search problems, so it reduces the computational complexity and does not need additional parameter pairing algorithm.A more efficient high-resolution localization and identification method for the moving sound sources is studied. This method consists of two parts:First, estimate the moving parameters, amend the coordinates of the virtual array in the moving direction. Second, generate the virtual array signals, achieve the virtual array steering vector, and obtain the spatial spectrum. The spatial resolution and the adaptability of the virtual array sparseness are both analyzed.This paper studies the experiments in both anechoic tank and Songhua Lake, and provides the experimental foundation for the engineering application of the noise sources localization and identification method based on the vector sensor array.