Study On Identifying Moving Sound Sources Based On Inner Product Correlation Characteristics

The control of noise source is always the concern of scholars at home and abroad,and accurate sound source location recognition is the precondition of noise control.The initial stage of the theory is mainly focused on the recognition of the stationary sound source from the development process of the sound source recognition method.With the deepening of the research,the diagnosis and tracking of the mobile sound source has gradually become the focus of the attention of the scholars.The core of this paper is the inner product correlation recognition of Cauchy Schwartz inequality,and the stationary point sound source recognition based on the inner product principle was used as the basis to study the moving point sound source,which was based on the virtual stationary point sound source recognition.By analyzing and summarizing its characteristics and scope of application,the method of using virtual moving point sound source to identify the moving sound source was proposed combined with the analysis of Doppler effect.The research of this paper was verified by simulation examples and experiments.The principle and process of stationary point source recognition based on the inner product correlation property were introduced,which laid the foundation for trajectory tracking of mobile source.A virtual sound source in the search space using the sound source frequency of the diagnosis is constructed,and the virtual and the actual sound source array sound pressure vector were obtained through the microphone array signal integration method,then the inner product of the two vectors was calculated.According to the inner product correlation characteristic,the virtual sound source position coincides with the actual sound source position when the inner product module value reaches an extreme value.The optimization algorithm was used to search the inner product module extremum to obtain the sound source position coordinates,and then the sound source intensity and its initial phase were calculated by the coordinates to complete the recognition of sound source.When the sound source moves,its sound pressure signal in a small period of time still has a certain correlation with the static sound source near the track of the sound source.The correlation and the stationary point sound source recognition theory were used to identify the moving source.The sound pressure radiation characteristics of mobile sound sources is analyzed,then the sound pressure signals were collected by the microphone arrays and the signals were segmented at equal time intervals,the static point source identification method is used to perform extremum search,then the position of the virtual sound source at the corresponding inner product extremum was obtained,and then the motion sound source trajectory over the entire time is tracked.Simulation examples and experimental results show that this method can accurately identify the low speed sound sources motion trajectory,which is not affected by the motion mode and has good anti-noise ability;but when the sound source speed is high,the trajectory tracking result has some error.T The influence of Doppler effect on the recognition of mobile acoustic source is discussed theoretically,then a method by using the virtual motion point source with higher correlation was proposed.The sound pressure radiation formula of the moving sound source and the inner product correlation characteristic were combined in this method,and the influence of the Doppler effect was corrected by the frequency search,the speed identification value was used to simplify the search process and then the movement segment by segment was completed.The simulation example and experimental results show that the method can correct the real time speed of the sound source and has a higher accuracy.Compared with the stationary point source,the method has higher recognition accuracy,and still maintains high recognition accuracy in the track tracking of medium and high speed sound sources.After using the principle of inner product correlation to further optimize the virtual source from static point source to the moving point source,the recognition method still preserves the flexibility of the former,and is not affected by the motion mode of the sound source,and can identify the sound source trajectory of the non uniform motion mode and the irregular path motion mode.