Tracking of 3D sound source location by particle filter with TDOA and signal power ratio

A new sound source tracking method in 3 dimensional space using state space modeling approach and particle filter as a state estimation method has been proposed. Not only TDOA (Time Difference Of Arrival) as our previous works including by the authors, but also using power ratio of two sound signals from microphone pair are effectively used in formulation of state space model. Simulational experiments consisting of 1) fundamental tracking with four microphones and 2) human head tracking in car cabin have been conducted with efficient performance in SNR 20 dB environment with about 3,000 number of particles.