TMPU: A Framework for Terrain Traversability Mapping and Planning in Uneven and Unstructured Environments

Abstract

In this paper, we propose an autonomous navigation framework (TMPU) for ground mobile robots in uneven and unstructured environments. The novelty of this framework is twofold: 1) The terrain traversability mapping module is proposed to build a low-resolution 2D (Two-Dimensional) occupancy map by plane analysis based on plane slope and terrain roughness. 2) A multi-resolution motion planning framework is proposed to obtain a high-resolution traversable corridor by A* low-resolution search. Then an improved global planning H-lattice planning is performed on the traversable corridor. Furthermore, extensive experiments are conducted to demonstrate the feasibility and safety of the proposed autonomous navigation framework. All test results show that the computing time consumption is reduced by up to 97.1%, compared with an advanced 3D navigation framework. (Supplemented video link: https://youtu.be/rvYoyR3imEk)