结合马尔可夫过程，提出一种概率论的声源定位算法，并给出了基于DSP的机器人实现。其中声源定位部分采用三个麦克风呈三角形放置，为减小由于噪声等引起的TDOA估计误差，采用马尔可夫过程计算时延，这样计算的时延会更可靠。该方案中的声源定位也属于一维定位，即只需知道声源的方向角-Combination of Markov process, a probability theory of sound source location algorithm, and give the robot based on the DSP implementation. Sound Source Localization in part one of the three microphones were used to place the triangle, in order to reduce the noise caused because of the TDOA estimation error, calculated using Markov process time delay, such delay will be calculated more reliable. The program of the sound source location is also a one-dimensional positioning, that is, only need to know the direction of the sound source angle