Research And ARM Implemetation Of Sound Source Location Techniques Based On Microphone Arrays

Nowadays, sound source location based on microphone arrays has a broad prospect of application in video conference, speech enhancement, hearing aids, hands free telephone and intelligent robot etc. So it has been received a growing interest and become a very hot area. In various source location method,time difference of arrival(TDOA) based on the double step method is the most widely used, which has simple principle, low calculated amoun, high precision, and can realize real-time in practice. The paper also focuses on this method.Firstly, simply introduces three kinds of existing sound source localization methods based on microphone array, and analyzes their advantages and disadvantages. Makes a deep study of near-field and far-field models of microphone array receives signals, and introduces the main differences between two models and their applications. Discusses three kinds of topological structures of microphone arrays commonly used, and gives the principles of microphone array design.Secondly, detailed discusses several common time delay estimation method. Makes a detailed formula deduction of geometric positioning methods based on time delay estimation. And analyzes their advantages and applications.Proposes an improvement model of microphone arrays, which is suitable for 2-d sound source location. And an accurate positioning formula of this model is given.Thirdly, based on EasyRAM2100 experimental plate, a two-dimensional and real-time sound source localization system is designed. The critical section of the system software is the double step sound source location method based on time difference of arrival(TDOA), the part of time delay estimation adopts cross-power spectral phase method, and the part of source localization adopts four yuan linear array localization formula.Finally, do a great amount of experiment test, the experimental results show that the hardware performance of this system is good, and the software localization algorithm has simple implementation, little computation and high accuracy. At the same time, the system has better antioxidant reverberation and anti-noise ability. So it can be used in real time positioning system.