Iterative Motion Planning for UAVs Based on Safe Airspace Skeleton and Kinematic Feasibility Constraints

Due to adverse environmental conditions, unmanned aerial vehicles(UAVs) are more easily affected than autonomous land vehicles(ALVs), making it more difficult for the control modules to follow the generated trajectories accurately. Besides, the limited payload capacity, which can only afford low-power onboard computers and low-precision sensors, leads to unreliable upstream localization and mapping. Therefore, robust motion planning systems are essential for safe navigation in environments with dense obstacles. In this paper, Iterative Motion Planning for UAVs Based on Safe Airspace Skeleton and Kinematic Feasibility Constraints is proposed, which consists of a safe airspace skeleton generation method leveraging kinematic feasibility constraints to enhance efficiency and solution quality. A motion risk query interface is proposed to replace traditional ESDF maps, which improves safety cost query efficiency. Additionally, hard and soft constraints are integrated by a dual-layer iterative heuristic module to optimize path planning safety and efficiency. Experiments are conducted in both simulations and real world, showing that the average planning time of the method is 0.17 seconds. Furthermore, all hazardous trajectory segments are eliminated and the process does not require parameter returning in different environments, proving its adaptability in various complex scenarios.