无人机路径规划与轨迹优化方法综述OA

The increasing demand for safe,real-time,and highly mobile autonomous flight in tasks such as inspection,surveying,logistics,and emergency response is driving the continuous development of motion planning technologies for unmanned aerial vehicles(UAVs).However,existing research has not yet systematically compared path planning and trajectory optimization methods,nor has it clearly elucidated the applicable scenarios of different methods in terms of optimality,computational complexity,and dependence on environmental prior knowledge.Therefore,a systematic review of the development of UAV planning methods from a unified perspective is provided in this paper.In path planning,it summarizes graph search,sampling-based planning,artificial potential field,and intelligent optimization methods,and analyzes their advantages and limitations.In trajectory optimization,it focuses on polynomial minimum-snap(fourth-order derivative),B-spline gradient optimization,time-optimal and numerically optimal control under explicit input and state constraints,and multi-objective energy efficiency and safety optimization frameworks supported by genetic algorithms,particle swarm optimization,and bee colony algorithms.Furthermore,the advances of deep reinforcement learning(DRL)in planning within complex and unknown environments are reviewed,its potential in terms of adaptability and robustness is assessed,and key challenges including sample efficiency and safety verifiability are addressed.Valuable references are thereby offered for future research directions,such as the integration of planning and optimization,DRL-based planning,and collaborative planning for multi-UAV.