针对面向城市、山区和密林等复杂环境的四旋翼无人机实时避障需求,提出了一种四旋翼无人机在线规避航迹规划方法。在分析四旋翼无人机空间运动的基础上,采用分段多项式对四旋翼无人机飞行航迹进行建模,引入距离代价、碰撞代价以及速度和加速度惩罚项,通过几何引导将初始航迹吸引到自由空间以构成规避航迹,结合航迹平滑估计对规避航迹的平滑度、动态可行性进行优化,从而实现四旋翼无人机的在线规避和航迹重规划能力。最后在复杂环境模型中,从计算时间、运算复杂度以及成功率等方面进行仿真比较。结果表明文中方法具有运算效率高、平滑度好、成功率高的特点,能够在未知环境、快速飞行过程中自动避障以避免碰撞威胁,更适用于四旋翼的在线规避航迹规划。

In order to meet the real-time obstacle avoidance requirements of quadrotor in urban, mountainous and dense forest environments, an online flight path planning method for quadrotor is proposed. On the basis of analyzing the spatial motion of the quadrotor, the segmented polynomial is used to model the flight path of the quadrotor. The distance cost, collision cost, speed and acceleration penalty terms are introduced. The initial track is attracted to the free space through geometric guidance to form an avoidance track. The smoothness and dynamic feasibility of the avoidance track are optimized by combining the track smoothing estimation, thus, the capability of online avoidance and flight path replanning of quadrotor can be realized. Finally in a complex environment model, the simulation results are compared in terms of the computation time, computing complexity and the success rate. The simulation results show that the method has the characterristics of high computing efficiency, good smoothness and high success rate. It can avoid collision threat automatically in unknown environment and fast flight, and is more suitable for online avoidance route planning of quadrotor.