Feature Recognition in Computer-Integrated Manufacturing

Article type: Research Article

Authors: Klauck, Christoph^a | Mauss, Jakob^b

Affiliations: [a] German Research Center for Artificial Intelligence, Inc. (DFKI), ARC-TEC Protect, D-67608 Kaiserslautern, Germany | [b] Labor für Künstliche Intelligenz (LKI), D-22765 Hamburg, Germany

Abstract: To integrate CA* systems like computer-aided design (CAD), computer-aided manufacturing (CAM), or computer-aided process planning (CAPP) with other applications in the world of computer-integrated manufacturing (ClM), one principal approach currently under development is based on feature representation. It enables any CIM component to recognize the higher-level entities—the so-called features—out of a lower-data exchange format, which might be the internal representation of a CAD system as well as some standard data exchange format. In this article, we present a “made-to-measure” parsing algorithm for feature recognition using the higher-level domain-specific representation language FEAT-REP—a representation language based on a (feature-) specific attributed node-labeled (feature) graph grammar (ANLFGG). The heuristic-driven chart-based bottom-up parser—called GraPaKL—analyzes attributed node-labeled graphs (representing workpieces) with a (feature-) specific attributed node-labeled graph grammar (representing the feature definitions) yielding a high-level (qualitative) description of the workpiece in terms of features.

DOI: 10.3233/ICA-1994-1501

Journal: Integrated Computer-Aided Engineering, vol. 1, no. 5, pp. 359-373, 1994