Based On Microphone Array Sound Source Localization And Tracking Algorithm Research

At present, the equipments based on microphone array emerge continuously, whether military or civilian products, The main functions of products have mainly sound source localization, and moving target tracking, but the product performance still to be improved. This dissertation studies the related algorithms based on microphone array, and makes the algorithms better. The main contents of dissertation are as follows:The characteristics of the speech signal, noise and reverberation are introduced. According to the distance between signal and microphone array, the near field and far field signal models are established. After the Detection Algorithm of various speech endpoints are analysed, the double detection algorithm combining time frequency variance method with energy method is proposed, and the simulation experiments are carried out which the effects of the structure of microphone array on the array signal are discussed.The sound source localization algorithms based on microphone array and MUSIC algorithm are discussed, and the target localization algorithm based on beam forming are majorly studied, which is the generalized cross-correlation algorithm. The improved time delay estimation algorithm based on generalized cross-correlation algorithm is put forward, and this algorithm has been verified by experiment simulations. The result shows that the stability and accuracy of target localization estimation are raised.The sound source tracking algorithms based on microphone array are studied. A relatively complete four kinds of source motion models are established by deduction, and applied in the interactive multiple model algorithm.An improved interactive multiple model algorithm is reseached, and particle filter algorithm is introduced.A speech signal model is established, and its endpoints are detected. Comparing performance of the multiple location estimation algorithms, various algorithms are used for target tracking, and the advantages and disadvantages of several tracking algorithm are summarized.