Particle filtering algorithms for speaker localization in a car environment

Human-computer interaction for in-vehicle information and navigation systems is a challenging problem. In this paper, we propose a robust system for speaker localization and speech enhancement for an in-vehicle speech dialog system by using microphone arrays. The performance of traditional localization methods drastically decline in a moderately reverberant environment. In this paper, we introduce Gauss-Hermite filter to integrate the current observation, effectively steer particles towards regions with high likelihood. The sound source localization (SSL) method based on the model of the precedence effect is integrated into the framework of speaker localization based on particle filters. The simulation indicates that the proposed algorithm is able to accurately track speaker in a moderately reverberant car environment.