Anomaly detection as creative insight
- The costs of undetected design problems are huge, especially for automobiles, and similar expensive, complex products manufactured in high numbers. If design issues are not discovered before production begins, the cost and impact of design changes increase dramatically. When design problems remain hidden all the way through manufacture, substandard products may reach the public, with severe consequences. This research began with a qualitative field study using grounded theory and ethnography to examine the activities of geographically distributed automotive design workers. The early detection of design issues was predicted to produce higher quality products and improved productivity. Research questions were chosen: What is an anomaly? What is the anomaly detection process? Anomaly detection was explored with an experiment in which test subjects (N = 65) reviewed images for anomalies. The primary finding was that anomaly detection is a significant human design activity. Specific definitions for anomalies and the anomaly detection process were developed. The steps in the anomaly detection process were identified. The differences between errors and anomalies were researched and described. An anomaly coding scheme was created. Anomaly detection was found to be learnable and teachable. Opportunities for the use of anomaly detection in design were revealed. It was found that anomalies are creative insights. The role of ambiguity in design representations was revealed. The study found a need for representations to be unambiguous and easily understood by users. The role of poor quality representations that prevent observers from detecting structural and functional anomalies was clarified.
|Type of resource
|electronic; electronic resource; remote
|1 online resource.
|Helms, Michael Ray
|Stanford University, Department of Mechanical Engineering
|Leifer, Larry J
|Leifer, Larry J
|Statement of responsibility
|Michael R. Helms.
|Submitted to the Department of Mechanical Engineering.
|Thesis (Ph.D.)--Stanford University, 2011.
- © 2011 by Michael Ray Helms
- This work is licensed under a Creative Commons Attribution Non Commercial 3.0 Unported license (CC BY-NC).
Also listed in
Loading usage metrics...