Applicaton Of Direction Arrival Estimation Of Vector Hydrophone Based On Neural Network

Compared with the scalar hydrophone,vector hydrophone has the advantages of small volume,good consistency,mature production processes and related technologies,so the vector hydrophone and array signal processing technology is one of the future research trend;As a representative of machine learning,artificial neural network is a self-learning,adaptive and parallel distributed information processing network structure.Therefore,it is a good choice to solve the problem of DOA estimation by using artificial neural network.This paper systematically studies the vector hydrophone array signal processing,through the construction of model hypothesis,feature extraction,network and simulation results validated the proposed method in terms of algorithm complexity and accuracy than other methods excellent.The main achievements of this paper are:In the third chapter,the DOA estimation of vector hydrophone based on particle swarm optimization and BP neural network is introduced.In this paper,the signal feature extraction method suitable for vector hydrophone is described in detail.The extracted signal features are used as the input of neural network,and the neural network model selects the PSO-BP neural network with good nonlinear fitting ability.The simulation results show that this method is not only convenient for feature extraction,and the precision and the running time of the algorithm are better than other algorithms.In the fourth chapter,the DOA estimation of vector hydrophone based on the generalized regression neural network is discussed.Firstly,the principle of optimization algorithm and generalized regression neural network are introduced.Then,we discuss how to optimize the generalized regression neural network.The simulation results show that compared with other methods,the network model constructed in this chapter has the advantages of simple structure,higher accuracy and higher running time.The vector hydrophone array signal processing is a complex problem and still need to work a lot of in-depth study in future.I wish this work can help to other researchers,and help the research and development of vector array signal processing technology.