Motion Planning for Climbing Mobility with Implementation on a Wall-Climbing Robot

Abstract

Future autonomous planetary explorers will require extreme terrain mobility to reach areas of interest, such as walled lunar pits and steep Martian rock layers. Climbing mobility systems are one proposed answer, requiring efficient and kinematically feasible motion planning for autonomous operation. Similarly, climbing planning is applicable to other micro-gravity situations requiring constant end effector contact with discrete handholds. This paper proposes a planning framework that poses kinematic climbing planning as a discrete optimal planning problem. Motion primitives are used to encourage large robot body workspaces and beneficial connections between climbing stances. A wall-climbing planner simulation is presented, along with implementation on a hardware demonstration testbed that successfully recognized, navigated, and climbed an arbitrary vertical wall.