Research On Underwater Target Recognition Technology Of Multi-static Sonar Based On Machine Learning

In modern sonar system,the automatic recognition of underwater target is always one of the key technologies.In recent years,with the rise of computer technology,the classification and recognition method based on machine learning has been applied in the field of underwater acoustics and achieved good results.Compared with the traditional single base active sonar and passive sonar,multi base active sonar has its own unique advantages,which has become one of the research hotspots of sonar technology in recent years.So,the technology of underwater target recognition of multi-static sonar is also worthy of research and exploration.The main research contents of this paper include:1.Numerical calculation method and Simulation of acoustic scattering field of underwater target transient signal.In this paper,the acoustic scattering characteristics and calculation method of underwater target are studied.The numerical simulation of acoustic scattering field of target is carried out by using the finite element and boundary element method in COMSOL software.Through the design and simulation modeling of the multi-static sonar system,the time-domain echo of the different direction nodes of the multi-static sonar target is obtained,which provides the data support for the follow-up research.2.Multi base sonar target recognition technology.According to the characteristics of the echo signal received by the node,the feature extraction method based on Hilbert Huang transform and wavelet packet entropy transform is used to analyze and extract the features of the target acoustic signal;the extracted feature vectors are sent into three designed classifiers for classification and recognition,and the recognition results under different conditions are compared and analyzed.The results show that the underwater target recognition technology based on machine learning has a unique recognition advantage,and gets a higher target recognition rate under low SNR.3.Waveform design of biological sonar.In the complex ocean environment,the emission waveform of active sonar will also affect the recognition performance.The signals transmitted by cetaceans,such as dolphins,have the characteristics of strong anti-interference ability and high positioning accuracy.In this paper,the bionic signal is designed according to the waveform of its transmitting signal;the bionic signal is used as the transmitting signal of the multi base active sonar to carry out the simulation experiment of target recognition;finally,the recognition results of two different transmitting signals bionic signal and LFM signal are compared and analyzed,which provides a new design idea for the waveform design of the active sonar.