Adaptive Event-triggered Control over a Shared Network

Adam Molin

Institute of Automatic Control Engineering

Technische Universitaet Muenchen, Germany

Friday, December 7, 2012, 12:00pm

Elliott Room 53-135

Abstract:

Under many circumstances event-triggered control outperforms time-triggered control schemes when resources such as communication, computation, or energy are sparse. In this talk, we analyze another

benefit of event-triggered control concerning the ability of adaptation. The system under consideration comprises multiple heterogeneous control loops that are closed over a shared communication network. Each subsystem is modeled as a discrete-time stochastic linear system. The design problem is formulated as an average cost Markov decision process (MDP) with unknown global system parameters that are to be estimated during execution. It is shown how techniques from distributed optimization and adaptive MDPs can be used to develop distributed self-regulating event-triggers that adjust their transmission rate to

meet a global resource constraint.

Biography:

Adam Molin received the engineering degree (equiv. to M.S.) in electrical engineering from the Technische Universitaet Muenchen, Munich, Germany, in 2007.  He is currently a Ph.D. Student at the Institute of Automatic Control Engineering, Technische Universitaet Muenchen.  His main research interests include event-based control, networked control systems, and stochastic optimal control.

For more information, contact Prof. Tabuada (tabuada@ee.ucla.edu)