Blind Nonparametric Estimation of SISO Continuous-time Systems
Augustus Elton (The University of Newcastle, Australia)
Rodrigo A. Gonzalez (Eindhoven University of Technology)
James S. Welsh (The University of Newcastle, Australia)
Tom Oomen (Eindhoven University of Technology, TU Delft - Team Jan-Willem van Wingerden)
Cristian R. Rojas (KTH Royal Institute of Technology)
More Info
expand_more
Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.
Abstract
Blind system identification is aimed at finding parameters of a system model when the input is inaccessible. In this paper, we propose a blind system identification method that delivers a single-input single-output, continuous-time model in a nonparametric kernel form. We take advantage of the representer theorem to form a joint maximum a posteriori estimator of the input and system impulse response. The identified system model and input are optimised in sequence to overcome the blind problem with generalised cross validation used to select appropriate hyperparameters given some fixed input sequence. We demonstrate via Monte Carlo simulations the accuracy of the method in terms of estimating the input.
help_outline