Research On Sound Source Location Based On Microphone Array

With the rapid and vigorous development of microphone array technology in various fields,research on sound source localization based on microphone arrays has also set off new waves of climax.One of the core issues to be addressed in acoustic source localization technology based on array signal processing research is direction of arrival estimation.Among them,beamforming technology plays a crucial role in the high-performance performance of microphone array systems.At present,how to achieve efficient,real-time,and accurate positioning in complex environments is still an urgent problem to be solved in this field.In this context,this article mainly studies array signal processing technology.Firstly,it studies the direction of arrival estimation technology of signals,and then focuses on the beamforming technology based on airborne sound sources.The main research contents of this paper are as follows:(1)A joint harmonic frequency and direction estimation(HFDE)method based on airborne acoustic target localization is proposed.When detecting airborne acoustic line spectrum signals,it is possible to lose a line spectrum of a target signal during the noise elimination stage,resulting in an error in detecting the fundamental frequency of a data snapshot signal.Only the frequency band in which the fundamental frequency of the signal is located is considered,and the utilization rate of harmonic frequency energy is low.The method proposed in this paper determines the fundamental frequency band through time-frequency graph analysis based on the line spectrum characteristics commonly found in airborne sound sources.Then,beamforming is performed with reference to the frequency doubling relationship between the fundamental frequency and harmonics of the airborne acoustic line spectrum signal.Finally,the harmonic frequency energy is superimposed on the line spectrum signal energy on the fundamental frequency to make the peak energy of the fundamental frequency line spectrum signal more prominent and achieve target direction estimation.This method avoids the fundamental error of the target signal,fully utilizes the harmonic energy of the target signal,and improves the accuracy of Direction of Arrival(DOA)estimation.Compared to the MVDR algorithm,the accuracy is improved by about 6.7% at a signalto-noise ratio of-20 d B.(2)A joint harmonic and azimuth orthogonal dipole estimation method based on airborne sound source target localization is proposed.Traditional direction of arrival estimation methods are limited in performance due to their small aperture,while differential microphone arrays have a small volume and frequency independent beam patterns,with dipoles being one of their basic types.Orthogonal dipole pairs can directly generate direction of arrival estimation for small aperture microphone arrays.Firstly,the signal subspace is estimated using the eigenvalue decomposition of the noise signal covariance matrix.Then,use orthogonal dipole pairs to obtain candidate values for estimating the direction of arrival in the first quadrant.Due to symmetry,the other three candidate values can also be obtained separately in the other three quadrants.Finally,traditional beamforming methods are used to select the estimated direction of arrival from the candidate values.In addition,the orthogonal dipole estimation method avoids spectral search and has higher computational efficiency,demonstrating good performance in direction of arrival estimation.Therefore,when applying the orthogonal dipole estimation method to airborne acoustic target localization,in order to better meet the estimation accuracy requirements of the orthogonal dipole estimation method,a joint direction of arrival estimation is proposed based on the harmonic frequency of the airborne acoustic line spectrum signal.When there is no reverberation and the signalto-noise ratio is 0d B,compared to the non-harmonic joint direction of arrival estimation,the accuracy is improved by about 6%.