Predictive system-level control

Optimising system performance by anticipating the system’s next operational use

Today’s high-tech embedded systems need to provide optimal performance under a wide range of inputs, either from the user or the environment. For example, a state-of-the-art printing system has to handle a large variety of paper types and print jobs. For the internal system design, each input category is represented by a separate operational mode.

During operation, a mode transition may cause a large disruption of the overall system performance. As a result, the transition between operating modes must be carefully managed.

The predictive system-level control method incorporates techniques that anticipate a system’s next operational use. It prepares the system’s transition from one mode to another as soon as a mode-switching event is foreseen. By iterative optimisation of system transitions, a higher overall system performance can be achieved.
 

Predictive system-level control




Predictive system-level control

Method details


Method description

  1. Create a system model with discrete modes and continuous states
  2. Develop a control algorithm that uses this model, and an input horizon to determine the best future system state and a smooth transition towards it
  3. Apply the control algorithm repeatedly to continuously guide the system to the best operational system state

Research project

Partners

Industrial partner:

  • Océ

Academic partners:

  • Delft University of Technology
  • Eindhoven University of Technology

Contact

Business manager:

Frans Beenker

Telephone:

+31 (0)88-866 5420

Email:

frans.beenker@tno.nl