A Sound Source Localization Method Based on the Active Perception of Mobile Robot

This paper applies human's active perception mechanism of external sound information to a mobile robot auditory system, an active sensing strategy based on dual microphones is proposed with regard to the issues of complex microphone array and positioning accuracy for traditional passive positioning and not adapting to the near-far field. Abstract the actual “action-evaluation” problem into the framework of the particle swarm optimization algorithm to obtain the current optimal position of the time and space information of the robot update sound. According to the optimal performance of the time delay algorithm, a particle optimization evaluation system based on geometric principle is established, and the evaluation of each position is guided by the principle of minimum mean square error. Establishing multiple sets of experiments for the home environment, the experimental results show that the robot can actively search for targets. Compared with the traditional single model, the positioning method based on active sensing can simultaneously adapt to the near, far field model, and autonomously reach the sound source position in space.