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:CUED Control Group Seminars
SUMMARY:Whither discrete-time model predictive control? -
Gabriele Pannocchia\, University of Pisa
DTSTART;TZID=Europe/London:20140306T150000
DTEND;TZID=Europe/London:20140306T160000
UID:TALK51233AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/51233
DESCRIPTION:In this talk I discuss and propose an efficient co
mputational procedure for the continuous time\, in
put constrained\, infinite horizon\, linear quadra
tic regulator problem (CLQR). To ensure satisfacti
on of the constraints\, the input is approximated
as a piecewise linear function on a finite time di
scretization. The solution of this approximate pro
blem is a standard quadratic program. A novel lowe
r bound on the infinite dimensional CLQR problem i
s developed\, and the discretization is adaptively
refined until a user supplied error tolerance on
the CLQR cost is achieved. The offline storage of
the required quadrature matrices at several levels
of discretization tailors the method for online u
se as required in model predictive control (MPC).
The performance of the proposed algorithm is then
compared with the standard discrete time MPC algor
ithms. The proposed method is shown to be signific
antly more efficient than standard discrete time M
PC that uses a sample time short enough to generat
e a cost close to the CLQR solution.
LOCATION:Cambridge University Engineering Department\, LR3
CONTACT:Tim Hughes
END:VEVENT
END:VCALENDAR