3D local planning for a forestry UGV based on terrain gradient and mechanical effort
Delano Oscar Do Rosario Lourenco, João Filipe Ferreira, David Portugal · Nottingham Trent University's Institutional Repository (Nottingham Trent Repository) · 2020
Planning feasible paths in 3D environments is a challenging problem, mainly in forestry environments due to, for instance, the rough and slippery terrain or the challenges for perception, as those caused by the high amount of trees, wind, and general unstructured nature of the environment. This paper presents a work in progress to propose an innovative method for 3D local planning in outdoor environments, to facilitate autonomous navigation of a forestry Unmanned Ground Vehicle. The proposed method builds on the ROS navigation stack integrating a module that analyses the gradient of the terrain to quantify slopes on the robot’s path, and taking into consideration its mechanical effort when planning paths to traverse.