Empirical and functional knowledge in an expert system for fault diagnosis

An expert system for car fault diagnosis is presented. The expert system's originality lies in the combined use of empirical and functional knowledge about a controlled apparatus. Another salient feature is the knowledge-base organization, which allows the two kinds of knowledge to be represented in a single database, so that the inference engines used (empirical and functional) may share the knowledge about the apparatus, and work together until no more information is found, or a solution is reached. The designer's major effort has been the integration of these two kinds of knowledge.<<ETX>>