Variant step size RRT: An efficient path planner for UAV in complex environments

摘要

Rapidly Exploring Random Tree (RRT) is a popular way for motion planning especially in high-dimensional environments. Efficient path planner is prerequisite for high-speed Unmanned Aerial Vehicles (UAV). In this paper, we proposed a path planning algorithm for UAV in complex environments based on RRT. Our contributions are mainly on three aspects. Firstly we proposed a novel RRT algorithm which can explore the complex environment more rapidly. It is achieved by adaptively changes the step size of the tree according to the location of obstacles. The proposed method also takes random points failed to pass the collision check process as indicators to speed up exploring. We proposed an optimizaion algorithm which can get the optimal path without time-consuming sampling. The simulation result demonstrates that our proposed can work efficiently.