Abstract
This project developed wall-climbing robots to carry non-destructive evaluation (NDE) devices for automatic data collection and surface/subsurface defect detection in concrete structures. By using simultaneous localization and mapping (SLAM), the robots can be located to allow for incremental 3D model reconstruction. An image dataset with pixel-level accuracy is collected and labeled to train a deep neural network (i.e., InspectionNet) with 12,000 iterations for each defect type. A ground penetrating radar (GPR)-based subsurface object detection method is proposed to implement visual inertial fusion for the estimation of the GPR pose, use a deep neural network module (i.e., DepthNet) to extract hyperbola features in the B-Scan GPR image, and predict dielectric to determine the depth of the objects. By synchronizing the robot motion and position information with GPR scan data, our new 3D GPR migration software can reconstruct and visualize the subsurface objects in 3D space. Finally, a robotic acoustic inspection system, involving a solenoid-based impacting mechanism and microphone, is proposed to detect delamination and voids through impact sounding. Through analyzing impact sounding signals, the power spectral density (PSD) method can highlight subsurface objects in a 3D map.
Recommended Citation
Xiao, Jizhong; Yang, Liang; Feng, Jinglun; and Hoxha, Ejup, "Final Report - Autonomous Wall-climbing Robots for Inspection and Maintenance of Concrete Bridges" (2021). Project AS-3. 1.
https://scholarsmine.mst.edu/project_as-3/1
Department(s)
Civil, Architectural and Environmental Engineering
Research Center/Lab(s)
INSPIRE - University Transportation Center
Sponsor(s)
Office of the Assistant Secretary for Research and Technology U.S. Department of Transportation 1200 New Jersey Avenue, SE Washington, DC 20590
Keywords and Phrases
Wall-climbing robot, ground penetration radar, impact sounding, machine learning, neural network
Report Number
INSPIRE-008
Document Type
Technical Report
Document Version
Final Version
File Type
text
Language(s)
English
Rights
© 2025 Missouri University of Science and Technology, All rights reserved.
Publication Date
February 28, 2021
Comments
Principal Investigator: Jizhong Xiao, Ph. D.
Grant #: USDOT # 69A3551747126
Grant Period: 11/30/2016 - 09/30/2024
Project Period: 03/01/2017 - 12/31/2020
The investigation was conducted under the auspices of the INSPIRE University Transportation Center.