Failure recovery in automated manufacturing cells has been attempted through a simple repetition of the previous operation, a finite state table look-up, and rule-based expert systems; however, failure recovery requires various sources of knowledge as well as a flexible and powerful control strategy. Further, the cause of failure should be identified lest the same failure occur over and over again. This article describes an opportunistic failure recovery approach adopting a multiple-blackboard architecture. A cause-based recovery strategy is proposed to demonstrate the opportunistic control aspects of the approach. The system has been implemented in the object-oriented language Smalltalk-80.
ASJC Scopus subject areas
- Theoretical Computer Science
- Computer Science Applications
- Computational Theory and Mathematics
- Artificial Intelligence