Tube-based distributed control for constrained systems
Event details
| Date | 30.11.2012 |
| Hour | 10:15 › 11:00 |
| Speaker | Giancarlo Ferrari Trecate (University of Pisa) |
| Location | |
| Category | Conferences - Seminars |
In recent years, man-made systems characterized by a large number of states or deployed over a wide area have triggered a renowned interest in decentralized and distributed control. While these problems have been thoroughly studied since the 70's for unconstrained systems, much less mature are methods for distributing Model Predictive Control (MPC) that is the standard for the regulation of constrained systems. Current research in distributed MPC tries to conciliate conflicting objectives such as weakening assumptions on system structure and constraints, minimizing real-time transmission of information between controllers and maximizing the operation range of the closed-loop system. In this talk we will illustrate a new distributed MPC scheme trying to fulfill most of these requirements. We will also consider decentralization of the control design process, that is desirable for easing revamping of control systems or addition/removal of subsystems. In particular, we will present a plug-and-play decentralized MPC method where the design of a controller for a subsystem involves only knowledge of subsystems physically coupled to it. Applications to the design of the automatic generation control layer for frequency control in power networks will be discussed.
Practical information
- General public
- Free
Organizer
- Colin Jones