Autonomous Wall-climbing Robots for Inspection and Maintenance of Concrete Bridges

Since 2002, the principal investor's (PI’s) group has developed four generations of wall-climbing robots for nondestructive evaluation (NDE) inspection of civil infrastructure. These robots combine the advantages of aerodynamic attraction and suction to achieve a desirable balance of strong adhesion and high mobility. They don’t require perfect sealing and can thus move on smooth and rough surfaces, such as brick, concrete, stucco, wood, glass, and metal. For example, Rise-Rover uses two drive modules to carry their middle compartment with payload up to 450 N. Ground penetrating radar (GPR)-Rover and Mini GPR-Rover are custom designed to carry a GSSI’s GPR antenna for subsurface defect detection and utility survey on concrete structures such as bridges and tunnels. The robots can also carry other devices such as impact echo and ultrasonic flaw detectors for bridge evaluation. To date, all the robots are remotely controlled to scan concrete surfaces. Even available, global positioning system (GPS) signals cannot accurately pinpoint the location of subsurface flaws. Therefore, a fast visual odometry method is proposed to estimate the pose and trajectory of a robot and in turn locate and mark the flaws. Features between consecutive images taken with a camera are detected and matched, and then fused with inertia measurement unit (IMU) data to achieve accurate motion control and feature registration/mapping. With this method, a robot with a camera can carry NDE devices, reach hard-to-access areas, take close-up pictures, scan the surfaces of bridge columns, girders, and decks (both top and bottom sides), record and wireless send NDE data to a host computer, log GPS coordinates, and record subsurface defect locations with respect to local landmarks at every sampling point for visualization and analysis. A robot can also be augmented with mechanisms and tools to perform maintenance actions such as cleaning, painting, and repairing. This project aims to develop motion control and localization methods to make wall-climbing robots a fully autonomous system with automated inspection process using various NDE devices and sensors, and design innovative mechanisms and tools and integrate them into the robots for maintenance actions. The focus of the first year will be to: (1) Design an embedded controller for GPR-Rover and develop methods to eliminate the potential interference between GPR signal and robot circuit, (2) Develop accurate motion control and localization methods that enable the robot to track grid patterns and take NDE samples automatically, (3) design a spray painting mechanism to mark defects, and (4) Evaluate the integrated robotic inspection system and the localization method using a GPR and a camera.

Project

Subject/Index Terms

Filing Info

  • Accession Number: 01645997
  • Record Type: Research project
  • Source Agency: Inspecting and Preserving Infrastructure through Robotic Exploration University Transportation Center
  • Contract Numbers: 69A3551747126
  • Files: UTC, RiP
  • Created Date: Sep 13 2017 4:37PM