A New Adaptive Filter Algorithm for System Identification using Independent Component Analysis

This paper proposes a new adaptive filter algorithm for system identification using independent component analysis (ICA), which separates the signal from noisy observation under the assumption that the signal and noise are independent. We first introduce an augmented state-space expression of the observed signal, representing the problem in terms of ICA, and then use an adaptive gradient descent algorithm to separate the noise from the signal. A local convergence condition is also shown. The proposed algorithm can be applied to the acoustic echo cancellation problem directly and some simulations have been carried out to illustrate its effectiveness.