Multi Target Detection And Parameter Estimation Based On Vector Hydrophone

The detection and parameter estimation of underwater acoustic targets are crucial issues in the field of underwater acoustics,and are of significant importance for the development of marine resources and maritime defense and security.Compared to pressure-based hydrophones,vector hydrophones can obtain more information and have higher processing gain.However,a single vector hydrophone has difficulty in resolving multiple targets.Nonetheless,due to its advantages of smaller size and simpler deployment compared to vector arrays,a single vector hydrophone still has great practical application value.In this paper,aiming at complex underwater environments and situations,we consider using a single vector hydrophone as a distributed measurement node.To improve the ability of a single vector hydrophone in multiple target detection,parameter estimation,and feature extraction,we studied the method of multiple target underwater acoustic signal detection,parameter estimation,and feature extraction based on a single vector hydrophone.We also investigated the method of parameter estimation and feature extraction of underwater multiple target radiated noise.The detection,estimation,and feature extraction results obtained can provide a foundation for subsequent target resolution and correlation in distributed systems.For the problem of detecting multiple air-drop targets entering the water,this paper uses the energy detection method to detect the signal and analyzes the influence of different forms of detection.For the multi-target situation,a method combining empirical mode decomposition and energy detection is studied to separate and reconstruct signals in each modal domain.After the signal is detected,the signals in each modal domain can be processed to obtain the spectrogram information of multi-targets in each modal order,which provides a basis for subsequent target localization and correlation.In order to separate and further extract features from the radiation noise signals of multiple targets,conventional beamforming in the expected azimuth direction is required for the received signals,and the azimuth estimation of multiple targets is particularly important.Among the multi-target azimuth estimation methods for single vector hydrophones,the method based on sound intensity is the most widely used.The average sound intensity method requires the multi-target signal to satisfy the time-domain separability condition,while the complex sound intensity method requires the multi-target signal to satisfy the frequency-domain separability condition.In order to handle complex radiation noise signals from multiple targets that cannot be separated in both time and frequency domains and extend the parameter estimation ability of a single vector hydrophone,this paper studies a multi-target parameter estimation method based on differential evolution algorithm,further analyzes the applicable conditions and parameter settings of this method,and simulates the azimuth estimation ability of this method under different source number conditions.In order to further separate the radiation noise signals of multiple targets,the obtained azimuth estimation results can be used as the expected direction for beamforming,which makes the output after beamforming have higher gain in the expected direction,suppresses the target signals in other directions to some extent,and is beneficial to extract target feature information.The spectral features of multi-target radiated noise include continuous spectrum and line spectrum features of the targets.This paper investigates the continuous spectrum feature extraction method based on spectral smoothing filtering and fitting,as well as the axis frequency extraction method based on DEMON spectrum.To obtain more comprehensive line spectrum information,the paper studies the segmented processing and fusion method of DEMON spectrum,and uses the obtained line spectrum information as weighting factors to normalize and sum the segmented DEMON spectrum,resulting in a fused DEMON spectrum of the target.Finally,the paper investigates the axis frequency-based line spectrum search and direction estimation methods.The obtained continuous and line spectrum information can provide a basis for subsequent multi-target resolution and correlation.