Evaluation of GCC-PHAT based Real-time Localization System in Noisy Environments

Abstract

Microphone array are designed to track speech souraces in noisy environments with improved speech quality. In our array designs, microphone array are arranged with various configuration with uniform and non-uniform. Previous methods to estimate speech source localization typically depend on conventional time-delay estimation techniques between two microphone pairs. However, such methods often ignore the ambient noise and reflections from the surroundings. The Generalized Cross Correlation (GCC) with Phase Transform (PHAT) as weighting function is employed in order to accurately detect, localize, and track speech sources. The methodology developed was evaluated for the various microphone arrays system under different condition of noise, to evaluate the of the system. The results demonstrate the accuracy localization, within a Root mean square error (RMSE) of 3°.