Windup properties of recursive parameter estimation algorithms in acoustic echo cancellation

The windup properties of a recently suggested recursive parameterestimation algorithm are investigated in comparison with a numberof well-known techniques such as the Normalized Least SquaresAlgorithm (NLMS) and the Kalman filter (KF). An acoustic echocancellation application is used as a benchmark for comparing theproperties of different approaches. The basic performance of themethod, both for white and colored input signal respectivelyappears to be similar to that of the KF and superior to the NLMS.When the energy in the input signal decreases, the algorithmperforms best of all compared estimation schemes. Once thesolution of the Riccati equation of the algorithm converged to auser defined point, it will stay there even though the inputexcitation is reduced. This explains the good anti-windupproperties of the method.