A fully integrated real-time multi-tasking knowledge-based system: application to an on-board diagnostic system

A flexible software architecture is required to allow problem solving in dynamic real-time situations. Such an architecture is described. It is based on a knowledge-based development system, SUPER, and a real-time multi-tasking kernel which gives to the knowledge-based system the ability to perform data input, interrupt handling, and temporal reasoning during the inference process. Methods are provided which allow the system to be responsive to important events and to focus attention dynamically. It is possible to generate a source code from the knowledge base. This code can be compiled and linked to external procedures. The application of the system to the development of an integrated real-time diagnostic system, embedded in a car, which computes its capacity to execute a maneuver safely, is described