MOBY-DIC main

From HYCON2
Jump to: navigation, search

MOBY-DIC Toolbox

The MOBY-DIC project is still in progress, and will end in December 2012. In the following, a brief description of the project and of the related toolbox is reported. More precise information will be inserted in more advanced phases of the toolbox development.

Description of the MOBY-DIC project

Central to the MOBY-DIC project is the development of PWA formalisms that bridge the gap between the control community and the embedded circuit community. In the past decades both communities conceived independently specialized PWA canonical forms tailored to their own purposes. As these goals are not geared toward each other, there is a need to develop canonical forms that are both control-oriented and circuit-oriented and can:

  • 
 facilitate efficient analysis and automatic synthesis of controllers in this canonical form;

  • easily be transformed into small-size and low-power circuits in a predictable manner.

Overview on MOBY-DIC toolbox

For the identified control/circuit co-design methodology to be successful and have an impact on industrial practice, the theoretical research achievements of MOBY-DIC will be immediately transferred into a complete computer-aided design tool chain. The MOBY-DIC project will adopt the principle that a good theory and a good tool strengthen each other in the following way. The embedding process is first represented through its mathematical dynamical model in a simulation environment such as Matlab’s Simulink (The Mathworks, Inc.). Based on such a dynamical model, the designer will be able to tune an optimization-based controller in simulation (either in PWA canonical form or not), until a satisfactory closed-loop performance is achieved. The availability of a complete, consistent, and highly automated design flow will drastically reduce the design effort and time of high-tech systems and infrastructure in various application domains. “First time right” designs come within reach and reconfiguration operations would be easy to conduct. In fact, possible changes of the underlying physics and topology of the controlled process (and, therefore, of its mathematical model) would automatically map in a new synthesized circuit without much effort.

For more information, please visit the Project Official web site

Personal tools