Bayesian kernel-based system identification with quantized output data
2015 (English)In: IFAC-PapersOnLine, ISSN 2405-8963, Vol. 48, no 28, 455-460 p.Article in journal (Refereed) Published
In this paper we introduce a novel method for linear system identification with quantized output data. We model the impulse response as a zero-mean Gaussian process whose covariance (kernel) is given by the recently proposed stable spline kernel, which encodes information on regularity and exponential stability. This serves as a starting point to cast our system identification problem into a Bayesian framework. We employ Markov Chain Monte Carlo (MCMC) methods to provide an estimate of the system. In particular, we show how to design a Gibbs sampler which quickly converges to the target distribution. Numerical simulations show a substantial improvement in the accuracy of the estimates over state-of-the-art kernel-based methods when employed in identification of systems with quantized data.
Place, publisher, year, edition, pages
Elsevier, 2015. Vol. 48, no 28, 455-460 p.
Impulse response, Linear systems, Markov processes, Numerical methods, Religious buildings, Bayesian frameworks, Gibbs samplers, Identification of systems, Kernel based methods, Markov chain Monte Carlo method, State of the art, System identification problems, Zero mean Gaussian process, Monte Carlo methods
IdentifiersURN: urn:nbn:se:kth:diva-195439DOI: 10.1016/j.ifacol.2015.12.170ScopusID: 2-s2.0-84988452886OAI: oai:DiVA.org:kth-195439DiVA: diva2:1046628
QC 201611142016-11-142016-11-032016-11-14Bibliographically approved