Expert system for autonomous handling of elementary assembly operations†

Publisher Website
Google Scholar
Abstract
This paper addresses the new conception of a real-time expert system for elementary two-arm assembly operations under the inclusion of operation-based vision. Every operation is represented by a local blackboard which contains a recognition part, a world model and a measurement part. The total set of these behaviour-oriented modules is supervised and scheduled by an operation monitor, which is again a blackboard. The local-knowledge sources of each individual blackboard receive sensor signals as input and can directly produce a reaction without going to a higher-level global-inference engine. This local, short connection between sensor signal and reflex is decided on the basis of compiled rules which have the form of production rules. The main types of these rules are regulation tasks depending on the operation and analysis rules for situation assessments. Global plans are produced as assembly graphs and transmitted in parts to the operation monitor as task specifications. A 3-D CAD modeller is used to support the object representation and recognition

This publication has 3 references indexed in Scilit: