Spoken Dialogue Interface Using Sound Field Control and Source Separation

摘要

For a robust automatic speech recognition of a hands-free spoken dialogue system without double-talk detection, we have proposed an interface to control the sound field at the microphone where the response sound of the system is not observed. In this paper, to give the method an ability to eliminate the redundant response sound following fluctuation of the room transfer function without double-talk detection, we introduce a semi-blind source separation technique which estimats independence of the observed signal and the known response sound. Ex-perimental results revealed that the proposed method improves robustness against fluctuation of the room transfer function.