改进卡尔曼滤波算法在语音信号处理中的应用

针对传统卡尔曼滤波算法在加性噪声影响下的语音信号处理中要先对被观测系统的状态进行统计估值和计算量很大的问题．提出了改进卡尔曼滤波算法。该算法不用预先估计噪声、驱动项以及语音模型参数，可直接得到增强的语音信号。通过仿真实验证明，该方法在减少计算量的同时，有效地消除了加性噪声，并能保持较好的语音可懂度。

Application and Research on Speech Signal Processing Using Improved Kalman Filter

Statistical estimation of the observation system state must be implemented, and there are a lot of calculation, when traditional Kalman filter is effected by additive noise in the speech signal process. The improved algorithm of Kalman filter is proposed for the problems above-mentioned. The enhanced speech signal is acquired directly without estimating noise, driving item and the parameters of speech model. The simulation results demonstrate the algorithm can not only reduce the calculation and eliminate a...