Degree Type

Thesis

Date of Award

1-1-2001

Degree Name

Master of Science

Department

Industrial and Manufacturing Systems Engineering

Major

Industrial Engineering

Abstract

Most current CAD systems have the tools to allow users to generate models using Boolean combinations of features. While current research has explored several directions for next generation CAD systems with a wider range of applications, there has been little work in providing assistance to the user for generating the models. The present research aims at using 3D object recognition techniques to recognize incomplete CAD models and thereby determine the user's intent to facilitate model development. A system has been developed to recognize complete and incomplete models belonging to a particular category for which the system stores a construction tree that describes the sequence in which features must be added in order to generate a model of the category. The construction tree of a model is analogous to the sequence of operations that would have to be performed to manufacture the part. The input CAD model is checked against the construction tree of the object in question using certain rules to obtain a confidence level representing the similarity of the input model to the object. The rules used by the system are classified as Shape Rules, Dimension Rules, Similarity Rules and Placement and Orientation Rules. The system recognizes models belonging to the category Gear, with sub-categories as Spur Gear (internal & external), Rack Gear and Straight Bevel Gear. Test cases are provided to display the system's competence and capability.

DOI

https://doi.org/10.31274/rtd-20201118-8

Copyright Owner

Aditya Abhinandan Ajmera

Language

en

OCLC Number

47046874

File Format

application/pdf

File Size

99 pages

Download

Share

COinS