Researchers from the Hefei Institutes of Physical Science of the Chinese Academy of Sciences have developed advanced robotic technologies to support the assembly and maintenance of future fusion reactors, addressing the critical challenge of operating heavy robotic arms with extreme precision in complex and hazardous environments.

They designed a novel robotic joint for heavy-duty robotics. By removing the sun gear from the traditional planetary gearbox, the researchers created additional space for power and control cables without compromising the joint's compactness. The joint's new three-stage transmission mechanism achieves an ultra-high reduction ratio of 13,806:1. It delivers torque of up to 139 kN·m with low backlash of 4.86 arc minutes. Tests confirmed the joint's strength, precision, and reliability, making it ideal for handling massive in-vessel components.

To solve the problem of precise peg-in-hole assembly in radiation environments, inspired by human hand–eye coordination, the researchers developed a deep reinforcement learning based method that integrates signals from a 2D camera and a force/torque sensor. The proposed system achieved  sub-0.1 mm accuracy, even without advanced 3D vision.

Regarding environmental perception, the researchers introduced Transformer-based Component Identification and Perception System (TCIPS), a model that segments 3D point cloud data into basic geometric primitives, such as planes, spheres, and cylinders. By capturing long-range spatial relationships and improving boundary detection, TCIPS enhances robot navigation in cluttered reactor settings.

Together, these innovations are a significant step toward building intelligent, heavy-duty robotic systems that can carry out complex, high-risk maintenance tasks in future fusion power plants.

Heavy-duty robotic arm joint and parameter optimization (Image by CHENG Yong)