| Compared with conventional sound pressure hydrophone, vector hydrophone can measure both sound pressure and velocity of sound field at the same time, it is also able to obtain directivity regardless of frequency to afford omnidirectional object bearing, which brings great potential for signal processing. Based on pressure gradient vector hydrophone, this thesis develops a model for multiple object direction estimation using the correlation between sound pressure and particle velocity of plane wave in isotropic noise field. Moreover, based on the presented model, a method for solving nonlinear correlated system of equation using Genetic Simulated Annealing Algorithm is proposed to implement multiple object direction estimation through a single vector hydrophone. Research works of this thesis include:Deduce principles and signal model of the inertia sensor and pressure gradient vector hydrophone, briefly introduce conventional methods of object bearing based on single vector hydrophone, and systematically analyze direction estimation error of single vector hydrophone.Using the correlation between plane wave pressure and particle velocity in isotropic noise field, develop a model for multiple object direction estimation.Analyze principles of Genetic Algorithm, Niche Genetic Algorithm, Simulated Annealing and Genetic Simulated Annealing Algorithm, compare searching performance of these algorithms, and analyze effects of Genetic Simulated Annealing Algorithm searching accuracy using different running and controlling parameters.Using Genetic Simulated Annealing Algorithm is proposed to implement multiple object direction estimation for solving a model for multiple object direction estimation. Conduct simulations using the proposed method in conditions of two narrowband and wideband sources, three narrowband and wideband sources. An improved method is also proposed to solve four wideband sources. Moreover, simulation research is done when source number is not a prior.
|
PDF Full Text Request