A comparative analysis of Kalman and gradient methods for adaptive echo cancellation

The suitability of different gradient methods for the adjustment of adaptive echo cancellers was investigated earlier. The investigations are extended to the application of the Kalman method and to a simplified version of it, which is based on the assumption that the adjustment signal is stationary and uncorrelated with respect to time and that the individual coefficients of the transversal filter are adjusted independently of each other. The method are simulated on the computer under realistic conditions and compared with the gradient methods investigated earlier with respect to the achievable echo return loss enhancement (ERLE) and the necessary implementation expense.