Real-time Path Planning Based on IRRTAlgorithm in 3D Environment

This paper presents a real-time path planning method for unmanned aerial vehicles(UAVs) navigation in 3D environment. In this study, 3D point clouds are processed into octomap, which is the octree-based 3D map representation of environment and is a probabilistic representation of occupancy including free and unknown areas. Rapidly-exploring random tree is improved by introducing the random probability and storing relatively optimal path with a certain number of iterations, called iteration rapidly-exploring random tree, to find a collision-free path in the free areas. Down-sampling and curve fitting by Bezier curve are performed to smooth path that satisfy the constraints of flight trajectory of UAVs. Experimental results verify the proposed algorithm and prove that it is valid.