Non-asymptotic uncertainty assessment of frequency responses using the LSCR approach

Abstract

In this paper, we develop methods for evaluating uncertainties in the frequency response of a dynamical system based on finitely many input-output data points. For this purpose, we extend the "Leave-out Sign-dominant Correlation Regions" (LSCR) algorithm by taking into account a priori information about the tail of the impulse response sequence of the true system. We develop a general algorithm which delivers a confidence region with a guaranteed probability for the frequency response at multiple frequency points. We illustrate the developed algorithm on simulation examples. © ICROS.