This study aims to address the significant challenges faced by live-line maintenance robots for power distribution networks, specifically the limitations of current operation methods in unstructured environments, which include pure teleoperation and fully autonomous operation, this work proposes a novel framework for a live-line maintenance robot system capable of overcoming the challenge.
The authors present a human-in-the-loop operation system in which human operators handle tasks that remain difficult to automate while also provide intelligent decision-making support for the robot’s autonomous operations. The autonomous operations are guided by a standard trajectory library, constructed using dynamic movement primitives (DMP), which models and reproduces demonstrated trajectories from expert demonstrations for diverse tasks. The effectiveness of the proposed system is verified through simulated live-line operations on a laboratory-built experimental platform.
The human-in-the-loop operation system effectively integrates human intelligence with robotic autonomy, thereby enhancing the efficiency and reliability of live-line maintenance.
This work introduces humanoid dexterous hands as end effectors for live-line maintenance robots, significantly improving their capability to perform complex tasks in unstructured environments. It further develops a human-in-the-loop operation system that leverages human expertise to support autonomous operations, thus addressing the limitations of pure teleoperation and fully autonomous operation. In addition, it establishes a standard trajectory library based on DMP, enabling the robot to model and reproduce expert demonstrations for various tasks.
