Research Of Moving Sound Source Localization Based On Microphone Array

In recent years, the use of microphone array signal source angle and position estimation has become a hot research direction. There are a large number of applications in audio and video conference camera automatically follow the sound localization, hearing aids denoising and enhancement, intelligent robot voice tracking and so on, but the trajectory of the moving sound source localization studies are still rare. Combining with the application widely based on arrival time delay difference(TDOA) and high resolution spectrum estimation(MUSIC) algorithm, research on trajectory of moving sound source location, estimate the sound source varies with time angle trajectory.First of all, studies three kinds of positioning methods, introduces the microphone array structure and signal model, introduces a uniform suitable linear array model for the movement of sound source localization.Secondly, analyze and compare several different time delay estimation methods, and the generalized cross correlation algorithm according to the different weighted function then do the simulation. And analyze the anti-noise performance of different weighted function, PHAT- GCC algorithm is discussed emphatically.Then, study the basic theory of the sub space of high resolution spectrum estimation(MUSIC) algorithm. According to complexity of the covariance matrix eigenvalue decomposition, use a value free algorithm to maintain the characteristics, spectrum estimation precision, improve the efficiency of estimating algorithm, and then do the simulation.Finally, according to the array signal model of moving sound source localization, use two algorithms which are generalized cross correlation(PHAT-GCC) and free of eigenvalue decomposition of the MUSIC algorithm to simulation, and then estimate the trajectory changes of source pitch in the process of angle moving. Then simulate and analyze the influence of the signal to noise ratio, the influence of sampling rate, and velocity of movement brought to the positioning system, for the frequency offset problem of mobile we brought into a correction.