Adaptive Noise Cancellation Using Kalman Filter for Non-Stationary Signals

Abstract

The present paper states with the Adaptive Noise Cancellation (ANC) of speech signal corrupted with additive Gaussian white noise. A new method is proposed based on adaptive Kalman filtering. The probabilistic approach of kalman filters over a packet-delaying network given the postpone distribution and decide the minimum required buffer length and algorithm has been used for estimation of unknown state variables within the system. Any work has been achieved for structures with fractional-order dynamics. The new techniques based totally at the kalman filter proposed within the beyond, function in steps: first the noise variance and the parameters of the signal version is estimated and secondly the speech signal expected. The strategies provided inside the paper gives an exclusive method consequently it does not require estimation of the noise variance. The noise variance estimation is always will become part of the kalman advantage calculation. The results of the application of Kalman filter on a non-stationary acoustic signal indicated that SNR of the ∼ 1.17 dB and MSE`0.032 can be achievable using Kalman filters and Kalman filter can be efficiently used for noise cancellation in place of other adoptive filters.