Bernoulli Filter for Super-Distance Speech Tracking in Outdoor Environments Using a Circle Microphone Array. (February 2019)

• Record Type: Journal Article
• Authors: Zhang, Yehui; Pan, Xiang
• Is Part Of: Journal of physics. Volume 1169(2019)
• Publication Date: 2019-02
• Subjects: Physics -- Congresses; 530.5
• DOI: 10.1088/1742-6596/1169/1/012030
• Languages: English
• ISSNs: 1742-6588

Abstract

Abstract: It is a challenge to track super-distance speech in an outdoor environment due to the existence of wind noise and interference sources. After the weak speech has been enhanced by beamforming carried out over a microphone array, Bernoulli filter is utilized to track the speech source. Bernoulli filter is one kind of Bayesian filters under the random finite set (RFS) framework, which can accurately estimate the presence and disappearance of the speech source. The outdoor experimental results have shown that the proposed method can effectively track the interesting speech.