Study On Principle And Method Of Broadband Underwater Acoustic Array Signal Processing

Broadband array signal processing is a main direction of underwater acoustic signal processing, but the broadband signal processing is more complex than that of narrowband. This dissertation is devoted to the study of several aspects of the broadband signal processing, concerning broadband coherent interference suppression, design of broadband array and DOA (Direction of Arrival) estimation of broadband source.The signal-phase-matching principle of three and multi sensors are proposed in this dissertation. By using these principles, the formulae of estimated spatial signal are derived. The physical meaning and the geometric interpretation of these principles are given. The performances of estimated signal are analyzed when the phase of expected signal and the amplitude of noise are fluctuant.The method of broadband coherent interference suppression is presented by using the signal-phase-matching principle. The formulae of estimated signal form coherent interference are derived. The sensor space that must be satisfied is derived. The performances of coherent interference suppression are analyzed in detail. The theoretical analyses and results of computer simulation show that the method has good performances in suppressing broadband coherent interference, while the DOA is unknown.A sample numerical design method for constant beamwidth beamformer is given. This method is easy to implement and has no restriction on array geometry and sensor patterns; Three methods for the design of broadband frequency invariant beamformer are proposed, which are the method of resampling signals received by the array, the DFT (Discrete Fourier Transform) interpolation method and adaptive synthesis method, respectively. Examples and performance analyses are given for each method; A simplified linear constrained minimum variance broadband adaptive algorithm that based on DFT interpolation method is presented. Compared with conventional broadband adaptive algorithms, the presented method can reduce computation work greatly.Broadband source location method is investigated by using signal-phase-matching principle, and simulations and experiments are carried out to show the validity of the method; Two high resolution algorithms of broadband signals are proposed, which are the DFT interpolation algorithm and the frequency invariant beam space algorithm. The proposed algorithms have no requirement for prior information of source locations, so the results of DOA estimation can be unbiased; A method of temporal-spatial averaging, which can improve the performance of DOA estimation for uniform linear arrays, is given. This method can be applied in the most narrowband or broadband eigen-structure based high resolution algorithms.