基于改进蒙特卡洛树搜索的无人机任务规划方法

For tackling the UAV mission planning challenge within complex 3D environments，the flight distance cost is assessed via a vertical-section terrain-following approach in mountainous topography characterized by digital elevation models，with thorough consideration of terrain impacts on cost estimation. An improved Monte Carlo tree search（IMCTS）method is introduced to derive mission planning sequences，resolving parameterization difficulties through normalized reward value intervals and harmonizing exploration-exploitation trade-offs. To expedite optimization of multi-segment continuous flyable trajectories for fixed-wing UAVs post-sequence acquisition，an integrated multi-segment Bezier trajectory optimization methodology is proposed. Simulation outcomes demonstrate that relative to conventional Monte Carlo tree search，greedy，and random search algorithms，IMCTS parameters remain impervious to decision space reward values，enabling faster convergence to near-optimal sequences. The integrated Bezier optimization framework efficiently generates high-order continuous trajectories adhering to dynamic constraints based on mission sequences.