Quadrupedal Legged Locomotion on Icy Regolith

As robots become prominent in planetary exploration and terrestrial missions, the ability to traverse challenging terrains is essential. Planetary-analogue environments such as icy regolith pose unique mobility challenges, where surface terradynamic properties strongly affect locomotion. Preliminary field tests with a quadrupedal robot revealed distinct failure modes: increased slippage on icy snow and greater sinkage on softer snow. Prior work on rough-terrain legged locomotion has largely emphasized learning- or perception-guided adaptations to unknown ground, with few demonstrations on ice or snow and limited insight into failure mechanisms arising from the unique mechanical behavior of such terrains. By contrast, snow/ice mechanics studies offer extensive penetrometer and material tests but are typically quasi-static and fail to capture the high toe impact velocities and dynamic responses encountered during locomotion. We address this gap by (i) measuring the penetration resistance and its dependence on speed in icy simulants and (ii) embedding these terradynamic models into terrain-aware gait adaptation. These terrain-informed strategies can improve locomotion efficiency and stability on icy terrains, advancing legged mobility in extraterrestrial and polar environments.