Efficient sound source localization method using region selection

Sound source localization systems in service robot applications estimate the direction of a human voice. Time difference information obtained from a few separate microphones is widely used for the estimation of the sound direction. The cross-correlation function is computed in order to calculate the time difference between the two signals. Inverse cosine function is used when the position of the maximum correlation value is converted to an angle. Because of the nonlinear characteristic of inverse cosine function, the accuracy of the computed angle is varied depending on the position of the specific sound source. In this paper, we propose an efficient sound source localization method using region selection. By the proposed approach, the region from 0deg to 180deg is divided into three regions and only one of the three regions is selected such that the selected region corresponds to the linear part of the inverse cosine function. Only the selected region is used for the sound source localization. Thus considerable amount of computation time is saved and the accuracy of the computed angle is improved. By simulations, it is shown that the estimation error by the proposed method is only 31% of that of the conventional approach.