| Sound source localization technology based on microphone array has been widelyused in video conference, mobile phone, digital hearing aids and so on. As a promisingtechnique, ARM has the advantages of high performance, low cost and small size.Therefore, achieving high accuracy sound source localization by using ARM has beenwidely studied. So far, the multiple signal classification algorithm is the most accurate, butit is difficult to be applied to ARM system for the high algorithm complexity. In this paper,the multiple signal classification algorithm is further studied and an improved method isproposed and achieved on ARM.Firstly, the sound source localization methods based on microphone array aregenerally introduced and the advantages and disadvantages of those are analyzed.Secondly, the model of the microphone array and the principle of the multiple signalclassification method are detailed indicated, respectively. And then the simulation resultsand discussion are given. Based on the simulation, an improved MUSIC algorithm, whichreduces the complexity of the conventional MUSIC algorithm by combining the maximumpower frequency location algorithm and the svdfree algorithm, is studied and successfullylocates the multi-frequency signal.Thirdly, the real-time sound source localization system based on ARM isdemonstrated by using the STM32F103RBT6processor. The hardware and the software ofthe system are respect introduced. The system consists of the sound sampling unit basedon microphone array, the analog signal preprocessing unit and the signal recognizing andlocating unit. Among the whole system, the improved MUSIC algorithm in the third unitis the most part.Finally, the designed system is tested in the environment with much noise. Theexperiment results show that the improved algorithm has better performance inlocalization. The estimation error and root mean square error can be controlled within asmall range, which enables the improved algorithm to be used in real-time system. |
PDF Full Text Request