BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//talks.cam.ac.uk//v3//EN
BEGIN:VTIMEZONE
TZID:Europe/London
BEGIN:DAYLIGHT
TZOFFSETFROM:+0000
TZOFFSETTO:+0100
TZNAME:BST
DTSTART:19700329T010000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0100
TZOFFSETTO:+0000
TZNAME:GMT
DTSTART:19701025T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
CATEGORIES:Signal Processing and Communications Lab Seminars
SUMMARY:Sigma-Points\, Cubatures and Rao-Blackwellization
in Recursive Bayesian Estimation - Simo Särkkä\, D
epartment of Biomedical Engineering and Computatio
nal Science Aalto University\, Finland
DTSTART;TZID=Europe/London:20110601T141500
DTEND;TZID=Europe/London:20110601T150000
UID:TALK31363AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/31363
DESCRIPTION:Although\, sequential Monte Carlo based particle f
ilters and smoothers\, in principle\, provide the
Bayesian solution to any dynamic estimation proble
m presentable in probabilistic state space form\,
they are not flawless. In particular\, as the stat
e dimension grows\, the required number of particl
es quickly becomes high. Furthermore\, plain parti
cle filters and smoothers cannot be used for infer
ring static or slowly varying parameters. One way
to overcome these problems is to resort to Gaussia
n approximations for the full state posteriors or
parts of the posteriors. Sigma-point and numerical
cubature integration based filters and smoothers
are a recently developed class of methods for robu
st computation of such Gaussian approximations. Ra
o-Blackwellization refers to partial closed form m
arginalization\, which can be used for avoiding sa
mpling of Gaussian or approximately Gaussian parts
of the state state\, or for marginalizing the sta
tic parameters of the model in closed form. In thi
s talk I will discuss the sigma-point\, numerical
integration and Rao-Blackwellization based filteri
ng and smoothing methods\, and present application
s of the methods.
LOCATION:LR11\, Engineering\, Department of
CONTACT:Rachel Fogg
END:VEVENT
END:VCALENDAR