Modeling systems based on noisy frequency and time domain measurements

Abstract

The Loewner matrix framework can identify the underlying system from given noise-free measurements either in the frequency, or in the time domain[1, 2]. This paper provides an analysis of the effects of noise on the performance of the SVD implementation of the Loewner matrix framework for different noise levels and proposes an improved approach which is able to identify an approximation of the original system even for high levels of noise. Moreover, for frequency domain measurements, our framework can handle systems with a large number of inputs and outputs while requiring small computational time and storage. © 2010 Springer-Verlag Berlin Heidelberg.