Home
Tags
No Tags
Development of a neocognitron simulator for GT
| Title: | Development of a neocognitron simulator for GT |
| Author: | Kulak, Fuat |
| Description: | Group technology is an approach to manufacturing that attempts to enhance production efficiency by grouping similar activities and tasks together, then reusing the results of previous activities in the execution of others found to be similar. The concept can be applied to a variety of activities such as design retrieval, purchasing, sales, and process planning. Traditionally, classification and coding has been used to implement group technology. In this thesis, however, a novel implementation using neural networks a technology noted for its pattern-matching capability, is used. This implementation focuses on an application to retrieve previously designed parts. The neocognitron, an artificial neural network paradigm, is selected after comparisons with other neural network families due to its characteristics of recognizing patterns without being affected by the shifts in the position and distortions. Two- dimensional representations of engineering designs are input to a neocognitron neural network to produce groups or clusters of similar parts. These representations, in their basic form, amount to pixel by pixel maps of the design, and can become very large when the design is represented in high resolution. A neocognitron neural network simulator using the C programming language was developed to implement the pattern matching engine. A pilot test was performed to validate this software. This system demonstrated the feasibility of training a neocognitron neural network first to recognize a generic pattern (representing a family), and then to recall a family of similar parts when queried with a part design. |
| Permanent Link: |
http://rave.ohiolink.edu/etdc/view?acc_num=ohiou1178129123
http://hdl.handle.net/2374.OX/13863 |
| Date: | 1994 |
Files in this item
| Files | Size | Format | View |
|---|---|---|---|
|
There are no files associated with this item. |
|||