Tid: 22 oktober 2007 kl 15.15-17.00
Plats : Seminarierummet 3733, Institutionen för matematik, KTH, Lindstedts väg 25, plan 7. Karta!
Föredragshållare: Thomas Schön, Avdelningen för reglerteknik, Institutionen för systemteknik, Linköpings universitet
Titel: An introduction to the particle filter and its applications
Sammanfattning: During the past 10 years the particle filter has become a very popular and powerful tool in solving the nonlinear state estimation problem in dynamical systems. Many variants of the particle filter have been suggested, and there is by now a huge literature on the subject. However, the basic algorithm is the same and all the different variants correspond to various design choices of this basic algorithm. We will in this seminar explain and derive the basic particle filtering algorithm in detail.
There will also be a thorough introduction to the underlying problem area, i.e., estimation in nonlinear and non-Gaussian dynamical systems. Several successful applications (such as GPS-free positioning of cars and navigation of fighter aircraft) of the theory are also discussed and some recent L(p)-convergence results will be given.