Thesis
Development of a Mobile Robotic System and Communication Framework Based on Ros2 for Vision-Based Detection of Damage to Nuclear Storage Containers
Washington State University
Master of Science (MS), Washington State University
2023
DOI:
https://doi.org/10.7273/000005347
Abstract
Nuclear storage facilities are sites of intense radiation, and surveillance of nuclear storage containers in these facilities is largely performed through manual inspection, which leaves humans performing the inspection vulnerable to the radiation. Robotic surveillance solutions in nuclear storage facilities for detecting damaged nuclear storage containers offer a viable alternative to reduce human exposure to nuclear radiation.
This work details the efforts made toward building such a solution through the development of a mobile robotic system (to autonomously scan container surfaces), a machine learning algorithm (to segment areas of damage on the scans), and a communication framework based on ROS2 (to integrate the two components together).
The capabilities of the system were tested in a laboratory environment using an array of surgical steel containers with varying degrees of damage on them. The containers were chosen and processed to best mimic the visual characteristics of nuclear storage containers. The system was successfully able to perform four tasks autonomously: 1) obtain scans of the container surfaces, 2) perform damage estimation on the scans using machine learning (ML), 3) transmit the scans
with the estimations for human operator verification and 4) retrain the ML algorithm on new data. The work establishes an end-to-end automated surveillance solution that allows the integration of multiple robotic agents, machine learning classification and human operator verification. It also provides an important resource for researchers developing algorithms for mapping and localization of mobile robots, eye-in-hand visual servo of robotic arms, and communication frameworks for integrating machine learning capabilities in robotic agents.
Metrics
4 File views/ downloads
23 Record Views
Details
- Title
- Development of a Mobile Robotic System and Communication Framework Based on Ros2 for Vision-Based Detection of Damage to Nuclear Storage Containers
- Creators
- George Sam Eralil
- Contributors
- John P. Swensen (Advisor)Nestor O. Perez-Arancibia (Committee Member)Ming Luo (Committee Member)
- Awarding Institution
- Washington State University
- Academic Unit
- Mechanical and Materials Engineering, School of
- Theses and Dissertations
- Master of Science (MS), Washington State University
- Publisher
- Washington State University
- Number of pages
- 127
- Identifiers
- 99901031138901842
- Language
- English
- Resource Type
- Thesis