Abstract: This paper studies the motion of a large and highly mobile six-legged lunar vehicle called athlete, developed by the Jet Propulsion Laboratory. This vehicle
rolls on wheels when possible, but can use the wheels as feet to walk when necessary. While gaited walking may suffice for most situations, rough and steep terrain
requires novel sequences of footsteps and postural adjustments that are specifically
adapted to local geometric and physical properties. This paper presents a planner
to compute these motions that combines graph searching techniques to generate a
sequence of candidate footfalls with probabilistic sample-based planning to generate
continuous motions to reach them. The viability of this approach is demonstrated
in simulation on several example terrains, even one that requires rappelling.