Vision-Based Interactive Path Planning for Robotic Bridge Paint Removal

Abstract

Since the 1970s, the toxic lead content in old paints has become a significant health concern. The common approach used in removing corroded paint for bridge paint maintenance relies on manual operation of blast nozzles by workers within a large containment enclosure installed around the blast area. The robotic bridge paint removal (RBPR) system was developed as a safe alternative method in the Construction Automation and Robotics Laboratory (CARL) at North Carolina State University. The goal of this project was to build and field test a prototype system that can increase the safety of workers and the public as well as protect the natural environment against toxic pollution. This paper presents a vision-based interactive control technology that was developed to support the telerobotic operation of bridge paint removal. The control method allows the operator to visually inspect the steel beam surface via a video camera and to interactively define the corroded area using a cursor. Subsequently, a motion path is automatically generated for a robot arm holding a blast nozzle to spot-clean the paint. The robot arm in turn moves the nozzle to the required position within the blast area and executes abrasive blasting. Field experiments demonstrated that the vision-based graphic interface provides an effective human-machine interface with sufficient flexibility for the robotic bridge paint removal in construction.