Echo cancellation without silent channel detection - Department of Information Technology

Recursive parameter estimation applied to the problem of echo
cancellation in a possibly silent channel is considered.

The application is used as a benchmark for comparing the properties of different
recursive parameter estimation algorithms, such as the Normalized
Least Squares Algorithm, the Kalman filter and a recently suggested
Kalman filter based anti-windup scheme.

The basic performance of the latter method, when the input signal
was white noise and colored noise respectively, was similar to
that of the Kalman filter. The tracking performance of the
algorithm was shown to converge exponentially
to the reference signal. When the energy in the input signal
decreases the algorithm provided the best parameter estimates.
This is due to the anti-windup properties of the method. When
the solution has converged to a user defined point, it will stay
there even though the input excitation is reduced.

Senior researchers

Alexander Medvedev, ((TorbjÃ¶rn Wigren))

Ph D student

Magnus Evestedt

Publications

A. Medvedev and M. Evestedt, "Elementwise decoupling and convergence of the Riccati equation in the SG algorithm", Automatica, available online since March 24, 2009.
M. Evestedt, A. Medvedev and T. Wigren, "Windup properties of recursive parameter estimation algorithms in acoustic echo cancellation", Control Engineering Practice, 2008, v.16, No. 11, pp.1372-1378.
A. Medvedev and M. Evestedt, "Elementwise decoupling and convergence of the Riccati equation in the SG-algorithm," 17th IFAC World Congress, Seoul, Korea, July 2008, pp. 10093-10098.
M. Evestedt and A. Medvedev, "Recursive parameter estimation by means of the SG-algorithm," 17th IFAC World Congress, Seoul, Korea, July 2008, 10081-10086.
M. Evestedt and A. Medvedev, "Stationary behavior of an anti-windup scheme for recursive parameter estimation under lack of excitation," Automatica, vol. 42, no. 1, pp. 151-157, January 2006.
Magnus Evestedt and Alexander Medvedev and TorbjÃ¶rn Wigren, "Windup properties of recursive parameter estimation algorithms in acoustic echo cancellation", IFAC World Congress, v. 1, Prague, 2005