Research on Lightweight Map-Guided Real-Time 3D Path Planning Methods

To address the challenges of high computational load and poor environmental adaptability in traditional path planning algorithms for Unmanned Aerial Vehicles (UAVs) operating in unknown and complex environments, this paper proposes a novel three-dimensional (3D) real-time path planning method guided by a lightweight map. The method enhances the traditional Tangent Bug algorithm by incorporating an artificial potential field (APF) model and a virtual obstacle mechanism to optimize its boundary point selection strategy, thereby significantly reducing computational complexity. Furthermore, it employs an octree to construct a probabilistic grid map, endowing the algorithm with global environmental memory capabilities and effectively circumventing local minima issues. Validation was conducted in 3D obstacle scenarios established on the Gazebo simulation platform, utilizing real-time perception data from LiDAR. The simulation results demonstrate that, compared to the traditional A* algorithm, the proposed algorithm, while rapidly establishing a lightweight environmental map, achieves a 96.58% reduction in replanning time at the cost of a 1.68% decrease in path optimality and a 5.6% increase in the maximum turning angle, significantly enhancing its real-time performance and environmental adaptability.