Research On Key Technologies Of Acoustic Positioning Based On Microphone Array

Acoustic signal processing based on microphone arrays has always been the focus of research,extensive acoustic monitoring,human-computer interaction,acoustic scene analysis and other fields.The spatial domain information of the signal is extracted after sampling.Based on this,the microphone array is studied based on endpoint detection,feature extraction,sound source enhancement,spatial optimization and sound source localization.The details are as follows:The problem of endpoint detection and feature extraction for single-channel pulsed acoustic signals is studied.Aiming at pulse sound,an improved endpoint detection method based on band variance and multi-band spectral subtraction is proposed,which can effectively reduce the influence of noise and reverberation on endpoint detection.Then a feature extraction method is proposed to obtain good time-frequency resolution.The problem of noise estimation and two-channel acoustic signal orientation enhancement is studied.A two-channel acoustic signal enhancement algorithm is proposed,which combines the coherence filter and the beamformer to suppress the noise by using the correlation of the signal and the spatial domain information respectively,so that the acoustic signal enhancement effect can be improved more effectively.The problem of planar grid array optimization is studied.A new planar array optimization method based on improved genetic algorithm is proposed.Simulation experiments show that an optimized array with good performance can be obtained.The problem of single source and multiple sound source localization is studied.The problem of sound source localization based on steerable beamforming and high-resolution spectral estimation is verified.An the self-organizing clustering algorithm is used to cluster the multi-frame positioning results to determine the indoor sound source number estimation and azimuth estimation in a short time.