Gait design for legged locomotion on granular slopes

Locomotion on granular slopes such as sand dunes remains challenging for legged robots due to the reduced shear strength and anisotropic yielding of granular media under gravity. To uncover the governing mechanisms, we experimentally measured granular resistive forces as a robotic leg intruded into laboratory sand with systematically-varied slope inclinations. Analysis of the normal and tangential force components revealed slope-dependent anisotropy in the granular resistance. By comparing applied leg forces to the local yield criterion, we developed a force balance model predicting a transition from sinkage-induced to slippage-induced failure as slope angle increases. The model further predicts that directing leg penetration angle upslope could reduce slippage on inclined sand surfaces. Preliminary field observations at White Sands National Park supported this prediction: on a 25 degree sand slope, a quadrupedal robot with upslope-penetrating gaits exhibited significantly-reduced slippage, improving locomotion performance up steep sand dunes. These findings demonstrate how physics-based insights can inform simple gait adaptations and enhance locomotion robustness on yielding slopes.