Knowledge representation and control structure based on three-dimensional symbolic skeletons for CAD/CAM integration

Abstract

A formalism for symbolic representation of threedimensional model and its use for knowledge representation and control structure are presented. A robust feature-based &sign (RFBD) approach has been developed to represent three dimensional objects and to provide meaningful geometric and topological properties for manufacturability evaluation. For knowledge acquisition, binary syntactic primitive pairs have been established for high level symbolic reasoning. Symbolic reasoning tables provide five stages for proper knowledge invocation. This hamework enables the system to reason about geomeeic shape based on syntactic pattern primitives and not on features. Both production rule systems and frames are used to represent the declarative and procedural knowledge sources in the EXCAST-1 .O (the expert system). Automatic meta-rule formation is proposed as a part of knowledge refinement process and the semantic data model is also proposed to aid various CAM applications for use in RFBD. The implementation of the algorithms and examples are provided.

Learning Theory as Applied to Mechanical CAD Training of Novices

“…Rodriguez, Ridge, Dickinson, and Whitman (1998) addressed the need to train mechanical CAD (Pro/E) users on proper modeling methodology by designing interactive computer sessions that guide the designer and allow him or her to explore what-if scenarios. Surprisingly, there is little work reported (Lang et al, 1991;You, Chu, & Kashyap, 1991) that applies to the learning of CAD skills in terms of either procedural or declarative knowledge.…”