The Random Walk Model Kalman Filter in Multichannel Active Noise Control

In order to adapt the Kalman filter (KF), the RLS algorithm or other, to single channel active noise control (ANC) the MFx structure is used. But, this is more complex in the multichannel case. The RLS and others have been adapted to multi-channel ANC. However, there has been no adaptation of the KF based on a random walk state space model. This is similar to the RLS but has some advantages. The RLS algorithm tries to invert the autocorrelation matrix of the reference signal, and can become unstable if this is singular. The same is not true for random walk based KF. In the paper we present an adaptation of the KF to multi-channel ANC.