Publication date: 10th April 2024
In recent years, the energy industry has ushered in an age of digitalization based on artificial intelligence (AI) technology. The associated AI techniques help, above all, to predict the potential of materials for future applications in large-scale industries. Then researchers conduct all the experiments to prove it. The work to date still keeps researchers in a work trap. Here we are the first to use a robotic arm to build a clean energy power generation unit, specifically solid oxide fuel cells (SOFCs), in the laboratory. A visual servo imitation learning system is used for robot-assisted cathode painting. We first use the visual information captured by an RGBD camera to identify and locate the cathode. An imitation learning framework is then applied to learn the painting path based on real human processes and transfer it to different conditions. Experimental results show that the robot helps to produce symmetrical cells and single cells, and both cells can be processed in the same way as by hand with 100% coverage and fine surface with high identity. This AI-driven robotic approach is more flexible in accomplishing multiple tasks in a creative laboratory with the assistance of a visual servo imitation learning system, compared to robots designed for limited tasks in industrial applications.