A Combined Approach to Localization for a Nuclear Cask Inspection Robot
Open Access
Author:
Beck, Jacob J
Area of Honors:
Mechanical Engineering
Degree:
Bachelor of Science
Document Type:
Thesis
Thesis Supervisors:
Dr. Sean N Brennan, Thesis Supervisor Zoubeida Ounaies, Thesis Honors Advisor
Keywords:
nuclear waste robotics inspection computer vision
Abstract:
As part of an initiative to develop methods to certify and confirm the integrity of nuclear waste storage casks, a team of researchers lead by Penn State is working to create an inspection robot to inspect dry-storage casks. This thesis focuses on the development of localization technologies that can be used inside the casks. The interior of these casks is accessible via narrow cooling channels, which present a low-feature, high-temperature and high-radiation environment. The work in this thesis explores various localization methods, develops visual simulations of the cask internals, evaluates identified methods using those simulations, and finally constructs and tests a hardware prototype of the localization system in a physical mock-up. We show that selected computer vision methods utilizing lasers and edge detection perform well in simulation, and could be applied for real-world localization with further work.
