Acoustic Source Localization Of Microphone Array Based On Improved GCC-PHAT Algorithm

With the improvement of people's living standards and the progress of science and technology,the application of service robots is also gradually promoted.For the service robot which is close to human daily life,it is necessary to realize humancomputer interaction better.Speech recognition has been applied to robot,and the development of sound source localization function will be an important research direction.It is under such a background that the method of sound source localization on service robot is studied in this paper.Corresponding solutions are proposed to improve their intelligence level.The main contents of this paper are as follows.(1)Several commonly used acoustic source localization methods are analyzed and compared.Finally,the method of acoustic source localization based on the arrival time difference of acoustic signals is chosen as the main line of this paper.(2)The signal preprocessing stage has been studied carefully.The basic signal m odel is classified.For the segmentation of speech segments and non speech segments,an effective method of speech endpoint detection” Frequency band variance method”is adopted,which can effectively remove the non speech segments and reduce the co mputational complexity.Since the general bandpass filter can only filter the noise out side the set frequency band,there is no alternative to the noise superimposed in the ba nd.Therefore,a speech enhancement method based on improved spectral subtraction is proposed.(3)Several time delay estimation methods are analyzed systematically,and their working principles are studied comparatively,Simulation comparison is also made.Fi nally,selected GCC-PHAT as the time delay estimation method,and improved its perf ormance in time delay estimation,the accuracy and reliability of it in the low SNR cas e of time delay estimation is improved,providing reliable delay value for the next pos ition estimation.(4)Two kinds of location estimation methods are introduced: geometric positioni ng method and objective function search method.Based on the traditional geometric p ositioning method presents a positioning method of seven microphone stereo cross arr ay,quickly narrow the scope of the location,effectively eliminate the ambiguity,impr ove the success rate of positioning.(5)After introducing the method of sound source localization based on the arriva l time difference of acoustic signal,the method is verified experimentally.Microphon e array to collect voice signals,the processor packaged the data,sent to the computer t hrough the serial port,using MATLAB software to analyze and deal with the data,an d ultimately get the sound source estimation position.The results show that the metho d used in this paper can locate the source of sound,and the accuracy is high,which ca n meet some practical needs.