City Research Online

Abstract

The thesis is concerned with the efficient interrogation of CAD data. CAD data finds use in diverse range of applications which necessitates extension and integration of the CAD data base. By an exhaustive categorization of such application requirements and analysis of various CAD techniques, it is shown that boundary data structures are the most suitable in CAD, CAM and advanced robotic applications.

Several boundary data structures have been proposed since the classic Winged edge data structure, these aimed at reducing the storage requirement and increasing information retrieval speeds. In this thesis methodologies are developed which enable us to discover compact and fast access time schemes and analyze and fine tune for individual applications. We demonstrate how the application of the optimality concepts can lead us to the discovery of more efficient data structures than popular data structures. All the boundary data structures proposed to date have been based on the underlying assumption that all the data resides in main memory. We show that in an integrated CAD environment (characterized by virtual a memory environment or a data base environment), these data structures are inefficient in both storage and time. We propose a new data structure shaped like A which is the most compact as well as more efficient in access time, under certain conditions of real memory and virtual memory. Experiments reveal a paradoxical phenomenon: access time increases with storage, violating the classic law of storage vs. time.

Recently non-manifold boundary geometric modeling has become popular to meet the growing needs such as uniform treatment of wire frame, surface and solid modeling and design by features. We introduce a uniform terminology and notation to distinguish and critically analyze several non-manifold boundary data structures. It is hoped to fulfill the need for a ready reference for the design of efficient boundary data structures. The other aspects dealt with are the validity and conversion of Boundary data structures.

To verify the concepts developed, in practice, a whole suite of fast algorithms have been implemented for model manipulation, visualization and data conversion.

Publication Type:	Thesis (Doctoral)
Subjects:	T Technology > TK Electrical engineering. Electronics Nuclear engineering
Departments:	School of Science & Technology > Engineering > Electrical & Electronic Engineering School of Science & Technology > School of Science & Technology Doctoral Theses Doctoral Theses

Export

Downloads