Terrain trafficability analysis for autonomous navigation: A GIS-based approach

This article presents a method of generating grid maps for autonomous navigation using a geographic information system (GIS). Besides developing ordinary map systems, a GIS can be applicable to the terrain trafficability analysis for an unmanned ground vehicle (UGV) by generating maps in which information about its spatial environment is encoded for path planning as part of autonomous navigation. For such maps, the most commonly used is a grid map over which a UGV’s paths can be optimized. As the UGV moves and its situation changes, grid maps need to be generated quickly in an automatic manner for new paths. First of all, a GIS is analyzed to identify meaningful data, layers and attributes, which are then related to the UGV’s navigation. For each grid in the grid map, every meaningful layer included is processed so that the set of values of attributes of the layer can be converted into the traverse cost. If multiple layers are included in a single grid, those traverse costs are fused and the result data is then replaced in the grid. Grid map generation is finished after all grids are filled with such costs. From test results using an experimental program, it is validated that the proposed method can be a viable tool for the timely generation of grid maps using a geographic information system for the UGV’s autonomous navigation.