BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Data-Enabled Predictive Control of Autonomous Energy Systems - Flo
 rian Doerfler (ETH Zürich)
DTSTART:20190503T103000Z
DTEND:20190503T113000Z
UID:TALK123631@talks.cam.ac.uk
CONTACT:INI IT
DESCRIPTION:We consider the problem of optimal and constrained control for
  unknown systems. A novel data-enabled predictive control (DeePC) algorith
 m is presented that computes optimal and safe control policies using real-
 time feedback driving the unknown system along a desired trajectory while 
 satisfying system constraints. Using a finite number of data samples from 
 the unknown system\, our proposed algorithm uses a behavioral systems theo
 ry approach to learn a non-parametric system model used to predict future 
 trajectories. We show that\, in the case of deterministic linear time-inva
 riant systems\, the DeePC algorithm is equivalent to the widely adopted Mo
 del Predictive Control (MPC)\, but it generally outperforms subsequent sys
 tem identification and model-based control. To cope with nonlinear and sto
 chastic systems\, we propose salient regularizations to the DeePC algorith
 m. Using techniques from distributionally robust stochastic optimization\,
  we prove that these regularization indeed robustify DeePC against corrupt
 ed data. We illustrate our results with nonlinear and noisy simulation cas
 e studies from aerial robotics\, power electronics\, and power systems.
LOCATION:Seminar Room 1\, Newton Institute
END:VEVENT
END:VCALENDAR