Adaptive model based parameter estimation, based on sparse data and frequency derivatives

Abstract

Rational functions are often used to model and to interpolate frequency-domain data. The data samples, required to obtain an accurate model, can be computationally expensive to simulate. Using the frequency derivatives of the data significantly reduces the number of support samples, since they provide additional information during the modeling process. They are particularly useful when the data is sparse and the samples are selected adaptively. © Springer-Verlag Berlin Heidelberg 2004.