Regresar

Expert system for an adaptive real-time control with supervisory functions.

Abstract:

It is known that it is possible to find a suitable combination of control and parameter estimation algorithms such that the overall system achieves a set of satisfactory performances, assuming that certain conditions for stability and convergence are satisfied. Some authors have proposed the development of a third level, the so-called supervisory level, which have two different goals 1) to check on-line the aforementioned conditions and 2) to take the adequate actions in order to assure a good functioning of the overall system when some of these conditions fail. In this paper we analyze the information set need at this level in order to take the appropriate actions. Practical considerations about the implementation of this level using classical languages and Artificial Intelligence tools are showed.