Intelligent control by intentional selection of reactions

Abstract

Research into intelligent control aims at the design of methodologies of control enabling intelligent capabilities to be reproduced in the machine. Four large families of approaches currently cover these capabilities. To combine their respective advantages the authors have developed a new approach seeking to confer greater autonomy on autonomous entities which should interact with a given environment and play a role in it. Ideas drawn from psychology and ethology have also been found useful in working out this approach, whose characteristics are described and whose functioning is illustrated from a simulated mobile robot environment.