A Particle Filtering Approach For Acoustic Source Localization And Tracking Using Microphone Arrays

Acoustic Source Localization and Tracking(ASLT) is to judge the speaker’s position information based on the speaker’s voice information, which is a very important subject in human-computer interaction, it can be widely used in multimedia systems, video monitoring system, video conference system and intelligent robot system, digital hearing aids and so on, and ASLT is the technology or the system core topics and techniques. For example, video conference system, ASLT technology can control the camera, the camera real-time pointing and tracking the current speaker; digital hearing aids, ASLT can automatically detect and automatically point the speaker speaker’s range, by way of achieving the speaker voice directional amplification and shielding the independent noise.Actually, there are some various noises, such as noises from air conditioning, lamp and unrelated vocals, and some of the desktop wall reflection caused by reverberation, which influence more or less the result of sound source localization and tracking system performance. Considering the noise and reverberation, this paper adopts a kind of steered beam forming particle filter method based on microphone array acoustic source localization and tracking. The method based on the particle filter framework, with steered beam forming (SBF) as the observation information, through the SBF function to construct the likelihood function, so as to realize the sound source tracking and positioning. Because of the standard particle filter particle degeneracy problem, particle degeneration can be prevented by setting resampling threshold. The experimental results show that, this method improves the sound source localization and tracking system in noise and reverberation ability.