Title :
A new kernel-based approach for system identification
Author :
Nicolao, Giuseppe De ; Pillonetto, Gianluigi
Author_Institution :
Dipt. di inf. e Sist., Univ. di Pavia, Pavia
Abstract :
We propose a new-kernel based approach for linear system identification. The impulse response is modeled as realization of a Gaussian process which includes information on smoothness and BIBO-stability. The corresponding minimum- variance estimate belongs to a Reproducing kernel Hilbert space which is given a spectral characterization and shown to be dense in the space of continuous functions. The approach may prove particularly useful in order to obtain reduced order models and assess the corresponding bias error in the context of robust identification. Several benchmarks taken from the literature demonstrate the effectiveness of the proposed approach.
Keywords :
Gaussian processes; Hilbert spaces; identification; linear systems; reduced order systems; stability; time-varying systems; transient response; BIBO-stability; Gaussian process; impulse response; kernel Hilbert space; kernel-based approach; linear system identification; minimum-variance estimate; reduced order models; Bayesian methods; Control systems; Gaussian processes; Hilbert space; Kernel; Linear systems; Reduced order systems; Robustness; Stochastic processes; System identification; Bayesian estimation; Gaussian processes; kernel-based methods; linear system identification; regularization; robust identification; stochastic embedding;
Conference_Titel :
American Control Conference, 2008
Conference_Location :
Seattle, WA
Print_ISBN :
978-1-4244-2078-0
Electronic_ISBN :
0743-1619
DOI :
10.1109/ACC.2008.4587206