A path-planning method for bridge cranes based on laser point clouds and signed distance field mapping

Automatic path planning is essential for improving the safety and efficiency of bridge crane operations in complex industrial environments. However, bridge crane path planning involves not only avoiding collisions with environmental obstacles but also satisfying the requirements of physical executability and operational efficiency in actual lifting tasks. To address these challenges, this study proposes a bridge crane path-planning method based on laser point cloud and signed distance field mapping. A multi-resolution signed distance field is constructed for efficient three-dimensional environment modeling, and a column-wise minimum signed distance field map is introduced to represent the vertical safety constraints of suspended-load motion. On this basis, a signed distance field-based physical A* algorithm is developed. It incorporates time, energy consumption, smoothness, and safety into a unified cost function and employs a multi-factor composite heuristic function to improve search efficiency and path quality. Experimental results demonstrate that the proposed method generates safer, smoother, and more physically executable trajectories than conventional planning methods, while exhibiting good adaptability and robustness under different payload conditions and operating scenarios.