Control Method of Wire Repairing Robot Based on Semantic Segmentation and Current Detection

Abstract The traditional repair of conductors in power systems, using manual work at height, carries a high level of risk. To ensure personnel safety and improve operational efficiency. Based on the development of a conductor repair robot using pre-stranded wires, the focus is on a semantic segmentation-based adaptive conductor arc sag control method, which enables the robot to adapt to conductors with different arc sags by segmentation identification of the conductor, offset calculation, and proportional-integral controller design to realize the robot hanging on the conductor. Subsequently, the winding gap is monitored according to the change of motor current, and a fuzzy controller is used to realize the robot winding operation. The robot prototype tests show that the robot is capable of adapting to the wire pitch and completing soft repair work.