A robot control architecture based on graph grammars and fuzzy logic

New concepts for robot control architectures are presented. Special emphasis is put on planning and monitoring for task level control tasks. New features introduced are behaviour switching, complete monitoring of plan execution and plan validity and rigorous explicit representation of activities and mutual dependencies within plans. The key techniques used to model decision making and plan modification are fuzzy logic and graph grammars. Fuzzy logic is used to guide planning and graph grammars provide the framework for expanding plan components. Moreover, a behaviour switching mechanism is proposed, which allows critical behaviours to interrupt or abandon a current less critical behaviour.<>