Particle Filters for Prediction of Chaos
2003 (English)In: Proceedings of the 13th IFAC Symposium on System Identification, 2003Conference paper (Refereed)
The use of particle filters for the prediction of time series arising from chaotic dynamical systems is explored. The specific dynamical systems considered are variations of the logistical map with an unknown parameter. This parameter is in the chaotic regime for these dynamical systems. The systems considered have both observation and process noise. The prediction algorithms studied are variations of particle filters which include a roughening technique. Cramer-Rao bounds for the prediction algorithm are developed and compared with Monte-Carlo simulations.
Place, publisher, year, edition, pages
Particle filter, Dynamical systems, Non-linear systems, Estimation, Chaos, Prediction, Cramer-Rao
Engineering and Technology Control Engineering
IdentifiersURN: urn:nbn:se:liu:diva-90298ISBN: 0080437095OAI: oai:DiVA.org:liu-90298DiVA: diva2:613660
13th IFAC Symposium on System Identification, Rotterdam, The Netherlands, August, 2003