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:Dynamic State Estimation using Dirac Mixture Appro
ximation and Directional Statistics - Igor Gilit
schenski\, Karlsruhe Institute of Technology
DTSTART;TZID=Europe/London:20131128T140000
DTEND;TZID=Europe/London:20131128T150000
UID:TALK48765AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/48765
DESCRIPTION:In many applications of stochastic filtering\, it
is of interest to consider \ninherently nonlinear
system models or inherently nonlinear domains such
as the \nsphere or the circle. This motivates the
development of nonlinear estimation \ntechniques
that are able to capture the nonlinearity and at
the same time are \nbased on a valid distributiona
l assumption.\n\nPropagation of continuous probabi
lity distributions through nonlinear \nfunctions m
ight be numerically burdensome and not solvable in
closed form. \nThus\, we propose a general framew
ork for approximating a given probability \ndistri
bution by another distribution. This is also of in
terest in other areas \nsuch as model predictive c
ontrol or information theory. Furthermore\, we \nd
iscuss scenarios where the Gaussian assumption mig
ht be inherently invalid \nwhich might happen for
estimation of angles or orientation involving hig
hly \nuncertain measurements or strong system nois
e. Thus\, filters based on circular \nand spherica
l distributions are proposed in order to handle th
is kind of \nproblems. Finally\, we will discuss t
he challenges involved in combining non-Gaussian d
istributional assumptions and approximate uncertai
nty propagation \ntechniques.\n\n\n*BIO*: Igor Gi
litschenski received his diploma in mathematics (m
ajor) and computer \nscience (minor) from the Univ
ersity of Stuttgart in September 2011. In \nNovemb
er 2011\, he joined the Intelligent Sensor-Actuato
r Systems (ISAS) \nLaboratory at the Karlsruhe Ins
titute of Technology (KIT)\, where he is working \
non his PhD within the research training group "Se
lf-organizing Sensor-\nActuator Networks". Igor co
-authored a paper\, which received the "Best Stud
ent Paper Award\, First Runner-Up" at the 16th Int
ernational Conference on Information Fusion.
LOCATION:BE4-38 (CBL Meeting Room)
CONTACT:Dr Ramji Venkataramanan
END:VEVENT
END:VCALENDAR