ABSTRACT
Inspecting concrete box girders is crucial for bridge maintenance. Challenges in accessibility and working conditions often hinder effective internal inspections, posing a serious threat to bridge operational safety. This paper proposes a solution using a lightweight and cost-effective robotic device for the automated collection of apparent internal defects in bridge box girders. Emphasizing the bridge environment and the lightweight nature of the device, we optimized mapping, path planning, and tracking schemes. Our robot, equipped with consumer-grade sensors, successfully mapped a hundred-meter-long box girder within a bridge at meter-level accuracy. Path planning covered the entire box girder, collecting defect information from the top and side walls.
