2D Topological Map Building by UAVs for Ground Robot Navigation
Yuqian Wang, Xuetao Zhang, Yisha Liu, Yan Zhuang · 2021 IEEE International Conference on Robotics and Biomimetics (ROBIO) · 2021
In air-ground cooperation, unmanned aerial vehicles (UAVs) are used to build a priori map of the ground environment from an aerial perspective, which is conducive to improve the navigation ability of ground robots. This paper proposes a 2D topology map building method from the air view, which can be effectively used for global path planning of ground mobile robots. To realize the 3D reconstruction of the ground from the air view, the 3D Euclidean Signed Distance Field (ESDF) is constructed incrementally from the Truncated Signed Distance Field (TSDF). The ESDF map occupies a large storage space, which is inconvenient to transmit to ground robots. To solve this problem, a lighter topology map is constructed on the basis of the ESDF map. Since ground robots can only move in the 2D plane, the 2D topological map is generated at any height from the ground to represent the topological structure of the ground robots working environment. The experimental results on the dataset and simulation environment shows the effectiveness of the proposed method.