This paper describes a novel approach based on the sequential Monte Carlo method for tracking of multiple sources overlapping in time. The algorithm is designed for multimicrophone applications and manipulates GCC-PHAT measurements in order to obtain robust likelihoods for all active sources. An experimental analysis was conducted on real data acquired with different microphone configurations. Results show the effectiveness of the proposed algorithm.
File in questo prodotto:
Non ci sono file associati a questo prodotto.