Abstract

Using UAV to create map of environment for further UGV path planning could give much benefits, because UAV has better field of view and could process big areas quicker. Different approaches exist to store resulting map (such as elevations maps, octomaps, etc.), but most of them are operating with point clouds at some step. This article shows practical way to directly use point clouds to plan traversable UGV path. It allows to skip discretization of space as usually done and utilize all data available. This work presents algorithms for point cloud based terrain assessment and global path planning with rapidly exploring random trees (RRT). Publicly available datasets of UAV-generated point cloud maps are used for experiments. Article gives quantitative estimation of performance of planning on point cloud map on CPU and practical realization details. Highlights-Proof practicability of UGV traversable global path planning directly on point cloud, generated from UAV; Shows experimental results on publicly available datasets; Gives quantitative estimation of performance of planning on point cloud map on CPU and practical realization details.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call