Three Dimensional Source Recognition And Location Baded On High-precision Beamforming Algorithm Combined With Umbrella Array Method

Base on three-dimensional focused beamforming with planar microphone array,there is a pseudo-acoustic source in the inverse aperture direction of the array,and it's weak to suppress background noise interference.On the other hand,the traditional beamforming algorithm(DAS)has a larger mainlobe width and lower spatial resolution.To break the limitations of the two problems of source identification,high-precision beamforming method based on umbrella array is presented for the 3D source identification and localization in this thesis.Simulation and experiment are conducted for three beamforming algorithms with two microphone arrays in the whole anechoic chamber.The main research contents in this thesis include:(1)Based on the mathematical model of three-dimensional focused beamforming with the umbrella microphone deduced by predecessor,the cross-spectral matrix algorithm with the umbrella microphone array mathematical model and the cross-spectral imaging function algorithm with the umbrella microphone array model are conducted.program of the cross-spectral matrix algorithm and cross-spectral imaging functions algorithm were completed,and verify the feasibility of the algorithm through the simulation and experiment.(2)The high-precision beamforming with the umbrella microphone array method is simulated.The advantages and disadvantages of the wheel array are measured through the mainlobe width of the X,Y,Z directions.The performance of the three algorithms of DAS,cross-spectral matrix and cross-spectral imaging function is measured by calculating the running time,main lobewidth,and the minimum distance between the two sources.(3)The results of the simulation analysis are verified and analyzed in the whole anechoic chamber.In the experiment,by changing the two parameters of the depth of the source and the frequency of the source,It is demonstrated that the mainlobe width in X and Y directions is reduced by using the cross spectral imaging function algorithm compared with the DAS algorithm,and the mainlobe width in the Z direction is only 0.4 times for the DAS algorithm.By introducing the absolute error,the relationship between the absolute error and the sound source depth and the sound source frequency is demonstrated.By calculating the running time of the algorithm,it is demonstrated that the performance of the cross spectral imaging function algorithm is better.In this paper,six beamforming methods of DAS algorithm,cross spectral matrix algorithm,cross spectral imaging function algorithm with wheel array and umbrella array are studied.Through simulation and experiment,taking the frequency of sound source and the depth of sound source as variables,it is verified that the cross spectral imaging function algorithm with umbrella array has better comprehensive performance in terms of computing time and spatial resolution.It provides a high quality method for the actual sound source identification.