Error recovery in intelligent robotic workcells

Abstract

This paper focuses on the requirements for achieving automated diagnosis and error recovery in a robotic assembly cell, Untended robust operation is the goal which is discussed, and the use of a knowledge-base oriented diagnostic/control system to support on-line error recovery is advocated. It is proposed that error recovery attempts should be made, not only based on information gathered by system sensors, but also by using additional cell-specific knowledge pertaining to recovery strategies. Recovery attempts are to be made even in the absence of complete diagnosis, in order to gather additional information, A control structure developed to implement this approach is presented.