Research Of Source Localization Algorithm Based On Microphone Array And Device Realization

Sound source localization technology based on microphone array is gradually applied in video conferencing, industrial noise localization, helicopter detection systems, impacting point measurement of military experiments and other practical environments. But the complexity of reality seriously restricts the development of sound source localization technology. Therefore, researching algorithm and device of sound source localization based on microphone array is of great significance.In order to improve sound source localization effects in indoor environments with complex noise and serious reverberation, this thesis mainly has accomplished the following jobs:Firstly, the far-field model and the near-field model are discussed, and the microphone being chosen is experimented. At the same time, the structures of common microphone arrays are introduced. Then, the localization algorithm of quaternary linear microphone array structure has been proposed. It was based on the linear array difference of sound distance, and it can effectively improve the localization stability.Secondly, the sound source localization algorithm based on TDOA is researched. As the peak is not obvious when using traditional basic cross-correlation delay estimation algorithm, the generalized correlation time delay estimation algorithm is mainly researched, and then the two algorithms are simulated and compared respectively in MATLAB, which is according to the characteristics of weighting functions. Further, the generalized correlation time delay estimation algorithm has been improved in order to eliminate the impact of noise and reverberation as much as possible.Meanwhile, the correlation function has been quadratic parabola interpolated under the low sampling rate so as to improve the spatial resolution of time delay estimation, the peak value search range of the correlation function is limited, which is according to the microphone spacing. The operation time of the algorithm is shortened. The zero pseudo peak caused by synchronous noise and synchronous reflected signal is eliminated. Enhance the credibility of the time delay estimation.Thirdly, the sound source localization device based on DSP is designed. In this thesis, TMS320VC5509A is chosen as control chip.The device can achieve four-way sound signal acquisition. Then, effective analog signal processing circuit is designed and real-time temperature compensation for sound velocity is carried out. At the same time, Software based on an improved GCC time delay estimation algorithm is programmed. At last, the PC interface based on the VB6.0 is designed, which can communication with RS232 serial port extended by DSP, and the locating results can be real-time display.Finally, software and hardware are debugged in the CCS environment. Then, a lot of localization experiments are tested in the laboratory environment to verify the feasibility of the algorithm. The average error of the positioning angle can be controlled at 10°. The average error of the positioning distance can be controlled at 0.2m. Meanwhile, this thesis analyses the error sources of the peak and the sound source localization as well as the overall system based on the test results.