一种基于两阶段的低空环境下固定翼无人机运动规划方法

This paper proposes a two⁃stage planning framework based on incremental heuristic search, aimed at ad⁃ dressing the dual challenges of dynamic obstacles and limited field of view encountered by fixed⁃wing aircraft during low⁃altitude flight in complex environments. Focusing on the safety⁃critical navigation issues of small fixed⁃wing UAVs in low⁃altitude environments, we present a comprehensive obstacle avoidance scheme suitable for typical complex scenarios is presented: a two⁃stage planning approach that integrates the incremental search algorithm life⁃ long planning A^∗(LPA) incorporating a lazy evaluation mechanism with motion primitives, enabling high⁃speed navigation in complex environments such as urban building clusters and jungle canyons. First, the Lazy⁃LPA algo⁃ rithm serves as the global planner, employing a lazy evaluation strategy to reduce redundant collision detection com⁃ putations, achieving millisecond⁃level three⁃dimensional path replanning in dynamic environments and providing to⁃ pologically optimal global guidance. Second, a motion primitive library that conforms to the dynamic constraints of fixed⁃wing aircraft is constructed, utilizing Dubins curves as heuristic guidance to generate smooth and feasible ob⁃ stacle avoidance trajectories locally. This scheme optimizes path search performance while ensuring real⁃time re⁃ sponse and stable operation in dynamic complex environments. Numerical simulations and field flight tests demon⁃ strate that the proposed two⁃stage planning algorithm achieves rapid trajectory replanning in complex environments, with computational efficiency significantly superior to that of traditional planning methods, thereby simultaneously satisfying the stringent requirements for high timeliness and adaptability to complex environments in low⁃altitude flight scenarios.