Robust parametric approach for impulse response estimation

A parametric technique for estimating the impulse response of a linear system using input-output observations in an outlier and distributionally uncertain environment is presented. The use of various cost functions for fitting the chosen output error model are discussed. By simulation, it is shown that the parametric approach based on the use of Huber´s function as a criterion for fitting the model is robust. It is also shown that even though the parametric model for the impulse response is only an approximation to the true impulse response, the estimates from this approach still outperform the nonparametric approach in the presence of contaminated noise and low SNR