Research on Information Fusion and Kalman Filter for Speech Enhancement

A novel method based on information fusion and Kalman filter for reducing the noisy signal's influence for processing the noisy speech signal in the time domain. It fused the speech data with two different filtering processes to get the combined error covariance matrix and the fused observation data, then process the fused observation data by the Kalman filter to achieve a better noise reduction effect. Therefore, the novel fusion Kalman filter meets the expected requirements of speech signal noise reduction, and the performance of its speech enhancement is better than the other different four filters.