A Visibility-Based Near-Optimal Planner for Robotic Inspection in Ultra High Voltage Centers

This paper proposes a visibility-based planner for autonomous inspection tasks in outdoor ultra-high voltage centers, using a mobile robot. Regular inspection of components in such installations is essential to prevent faults and avoid power blackouts. The proposed system enables users to specify particular components for inspection. Based on this input and the environment’s map data, our algorithm suggests the most efficient inspection path, identifying the most traversable route across the selected components while maximizing optical and infrared data capture. We decompose the task into smaller sub-problems, addressing each individually. First, we perform a visibility-based waypoint selection and geometric path planning task, both executed offline as pre-processing steps. Using this data along with the user’s input, we then frame the problem as a Generalized Traveling Salesman Problem and solve it using the GLNS solver, which demonstrates superior performance and suitability for our problem type compared to other state-of-the-art methods. Experimental benchmarks indicate that our system consistently identifies optimal solutions tailored to our specific inspection use case.