Sound source localization for robot auditory system using the summed GCC method

Abstract

A variety of methods for sound source localization applied to robot auditory system have been developed. Most of them mainly depend on the time difference of arrival (TDOA) between microphones because of light computational load and easy application. Generally, possible region to localize the source, whatever localization methods are used, depends on the number of microphones. In case of the localization method based on TDOA, minimum four microphones which donpsilat all lie in the same plane are needed to estimate the source direction in 3D space. However, the new approach based on the summed GCC method can estimate the source direction in 3D space utilizing three microphones only and the platform effect. Because microphones for the robot auditory system are usually installed on the outer robot platform the proposed algorithm is quite suitable for robot applications. Difference between mapping functions caused by robot platform makes the source localization in 3D space with three microphones only be possible. We have shown a case where the sound source localization in some restricted region of the 3D space is possible by using the proposed approach through the ideal simulation.