An Efficient Trajectory Planner for Fixed-Wing UAVs via Improved A*-Guided Initialization

Abstract

This study proposes an efficient trajectory planning method for fixed-wing unmanned aerial vehicles (UAVs) based on an improved A*-guided initialization strategy for complex obstacle environments. An improved three-dimensional A* method is developed to generate smooth and collision-free global paths with enhanced search efficiency and path quality. The generated paths are then converted into dynamics-informed state and control trajectories, which serve as initial guesses for trajectory optimization. Simulation results under different obstacle complexities show that, compared with conventional linear interpolation initialization, the proposed method significantly improves solution success rate and computational efficiency, while producing more reasonable obstacle-avoidance strategies, smoother state variations, and more coordinated control inputs.