Autonomous Location and Obstacle Avoidance of Inspection Robots Based on Multi-modal Information

Abstract

Currently, manual inspection methods are still the main means of inspecting overhead transmission lines, which suffer from issues such as high labor intensity, low inspection efficiency, and significant operational risks. This paper focuses on researching the self-positioning and autonomous obstacle avoidance of patrol robots in the inspection operation environment. By analyzing the operational environment of power inspection work and combining the point cloud map constructed in the previous self-positioning system, a 3DVFH-based obstacle avoidance path planning method is proposed. An autonomous obstacle avoidance system for Flying inspection robots is developed. To address the problem of large-scale point cloud data input from the self-positioning system, voxel down-sampling is employed for point cloud compression, and the point cloud around the robot’s specific area is processed into a voxel octree map. Simultaneously, in the process of searching for flyable path points in the algorithm, voxel weight calculation is simplified to ensure that the system can generate smooth obstacle avoidance paths between the searched path points.