| Nowadays,with the continuous development of sensing technology,signal processing and detection technology,the need to obtain directions of objects through sound is increasing,and the system based on sound orientation is widely used in the noise processing of vehicle engines and the “hearing systems” of the intelligent robots,etc.,together with which the sound source orientation techniques based on microphone arrays are also evolving.Compared to traditional single microphones,the microphone array can collect more information when collecting sound.In general,humans use two ears to determine the orientation,otherwise when the machine is used for orientation,more microphones can be used to compensate for the high complexity in the human auditory system,resulting in higher resolution in space.This paper mainly analyzes the main application of microphone array system to collect voice information,the basic types and characteristics of microphones,and analyzes the preprocessing of microphone signal acquisition,model establishment,and the structure of microphone arrays,and shows the derivation process of the number of array elements of the microphone array,determining the number of array elements.Then,an improved sound source orientation method based on the generalized weighting function of time difference of arrival(TDOA)estimation is proposed.Through the simulation of MATLAB and field test analysis,this method is suitable for far-field and near-field sound sources.Compared with the traditional weighted function,it has higher orientation accuracy and is easy to realize real-time system.The method is applicable to both far-field and near-field sound sources,verified by an experimental platform,and data analysis is performed as well.On this basis,the paper completed the construction of the system,tested the hardware and software system,and verified the functionality.The experiment proves that the improved algorithm has certain guiding significance and meets the requirements of this design. |
PDF Full Text Request