Map Construction Fusing Environmental Information and Motion Constraints

Abstract

Robots are expected to replace humans in performing scientific expedition tasks, which greatly reduces the essential impact of human basic survival on scientific expedition activities, and improve the space-time coverage of scientific expedition activities in polar regions. However, the polar regions have complex topographic features, such as the edge of the glacier on the Qinghai-Tibet Plateau is steep and the inland is relatively flat, which undoubtedly brings great challenges to realize the safe and autonomous navigation of unmanned ground vehicles. For glacier area of the Qinghai-Tibet Plateau, accurately converting environmental information into the map that robots can understand is an important part of path planning, that is, mapping. This paper proposes a method of constructing an occupancy grid map for large-scale point cloud data, taking into account the movement constraints and terrain information of the mobile platform. By dividing and rasterizing the sub-point cloud, the state of the raster can be judged according to the normal, elevation and other information. Finally, we carried out path planning based on the A* on the raster map.