关键词: 动态环境, 碰撞回避, 运动规划, 速度障碍物, 可达速度集

Abstract: In three-dimensional space,how to conduct motion planning when unmanned aerial vehicles(UAV)facing moving obstacles is an interesting research direction.In dynamic environments,traditional algorithms based on the speed obstacles mainly aim at two-dimensional robots,which realizes obstacle avoidance actions by selecting velocities from the reachable velocity set out of the collision velocity set.This paper generalizes algorithms based on the current positions and velocities of UAV and obstacles to calculate the set of velocities that will cause collisions.According to the maximum speed and maximum acceleration of UAV,the velocity space that can be reached at the current moment is restricted.Different strategies are formulated for the needs of various scenes and are selected in a specific method in this subtraction set,so as to avoid obstacles in a three-dimensional scene with specific requirements.Aiming at the scene where UAV abstracted as spherical travels to the destination by avoiding spherical obstacles in 3D space,this paper verifies the obstacle avoidance algorithm by combining C++ and blueprint programming.It captures the movement trajectory of different strategies and records the corresponding consumption time,which demonstrates that the proposed algorithm can effectively complete the dynamic obstacle avoidance task of UAV in three-dimensional space.

Key words: Dynamic environments, Collision avoidance, Motion planning, Velocity obstacle, Reachable velocity set