In: Records of the Conference on Proceedings of AI, Simulation and Planning in High Autonomy Systems, 1990, Tucson, AZ, USA, pages 80-88. Piscataway, NJ, USA: IEEE Service Center, 1990.
Abstract: A hierarchical modular structure is presented for real-time control and monitoring of flexible manufacturing systems (FMSs) in which a clear distinction between processing of normal and unusual states is made. It is shown that this approach furthers autonomous diagnosis and recovery, because it makes it possible to integrate two techniques well suited in their own domains: Petri nets for the specification of normal sequences of control and AI techniques for dealing with diagnosis problems. The implementation techniques proposed are similar to rapid prototyping by execution of the specification.
Keywords: flexible manufacturing system (control); recovery (of systems); artificial intelligence; rapid prototyping; production rules.