Outliers Based Double-Kalman Filter for Sound Source Localization

During the course of sound source localization, the localization system often misoperates because of the environment noise, sensors´ failure and other reasons. And there are outliers in the measurements because of the clutter environment. To solve this problem, this paper present a new Kalman filter based on outliers for sound source localization. By adaptively adjusting the measurement noise covariance based on the two results of Kalman filter with the different steps, the presented filter can detect the outliers and reduce the their influence. At last ideal effects are obtained in simulation experiments.