Towards intelligent execution supervision for flexible assembly systems

Abstract

Research results concerning the error detection and recovery in robotic assembly systems, the key components of flexible manufacturing systems, are presented. The approach to the integration of services and the modelling of tasks, resources and environment is described. A planning strategy and domain knowledge for nominal plan execution and for error recovery is presented. A supervision architecture provides, at different levels of abstraction, functions for dispatching actions, monitoring their execution, and diagnosing and recovering from failures. Through the use of machine learning techniques, the supervision architecture is given the capability for improving its performance over time.