Fig. 2
Left: a simulation environment was developed by the research team to replicate the real-world experimental setup in the measurement chamber. The figure shows the simulator’s visualization, where the robot manipulator is in its initial configuration with the compliant paint-roller attached. The simulation also includes the positions of the paint container and dry wall boards (shown in red). The robot’s motion is first programmed and tested in simulation for safety and effectiveness before deployed in the real setup. Accurate-enough reproductions of the geometries of the robot, the paint roller, the paint bucket and the walls, as well as corresponding software, are needed to produce motions that a avoid undesirable collisions, and b result in contact between the roller geometry and the target wall. Right: the figure shows the 3D digital model of the specially designed compliant paint-roller, which was attached to the robotic arm, right next to the real one. The real system was constructed from 3D printed components based on the digital model. The key feature of the paint-roller is that it has an internal spring-loaded mechanism—highlighted in the digital model—which provides compliance and robustness to positioning errors. This makes it possible to use the robot for the intended purpose without the need for expensive sensors and a complex sensor-monitoring process
