Knowledge representation approach based on concurrent objects: application to a “robot-vision” system

Knowledge-based systems, intended for applications in the field of robot-vision, involve particular constraints which are due to the large amount of data to be processed, on the one hand, and the complex, nonhomogeneous and evolutive nature of the information conveyed by such data, on the other. To solve this issue while keeping away from the classical monolithic approaches to representation, our approach is based on object formalism. The model proposed offers at the same time: representation structures adapted to the specification of the properties of manipulated knowledge and to the description of the various link types that can be supported by this knowledge; and a formal frame for expressing distributed activities. In such a model, any “robot-vision” system can be projected in the form of a modular decomposition of independent processes which communicate with each other through the message-transmission mechanism