Search papers, labs, and topics across Lattice.
This paper introduces HumoSlope, a two-stage physics-guided framework for enhancing humanoid locomotion on steep sloped terrains, addressing the limitations of model-free reinforcement learning in such environments. The first stage establishes a balance prior using a slope-adaptive Zero Moment Point (ZMP) regularizer, while the second stage employs the Biomechanical Slope Gait Adapter (BSGA) to dynamically adjust center-of-mass height and limb coordination based on terrain geometry. Extensive Sim-to-Real experiments show that HumoSlope significantly improves posture stability and enables continuous traversal of slopes up to 62.7%, demonstrating the effectiveness of integrating physics principles into locomotion strategies.
Humanoid robots can now traverse extreme slopes blindfolded, thanks to a novel physics-guided approach that prevents posture degeneration.
Model-free reinforcement learning has enabled impressive humanoid locomotion; however, control on steep slopes remains largely unexplored. Unlike flat or discrete terrains, sloped terrains impose a persistent gravitational bias that demands simultaneous stability and posture control. Consequently, under generic reward formulations, policies can converge to slow, conservative low-center-of-mass (CoM) crouched gaits. In this work, we propose a novel two-stage physics-guided framework, dubbed HumoSlope, dedicated to robust humanoid locomotion on diverse sloped terrains. Specifically, Stage I establishes a terrain-consistent balance prior by introducing a slope-adaptive Zero Moment Point (ZMP) regularizer evaluated directly on the local inclined support plane rather than a world-horizontal reference. To prevent the resulting policy from defaulting to a crouched posture, Stage II introduces the Biomechanical Slope Gait Adapter (BSGA). Utilizing extracted macroscopic terrain descriptors as privileged, training-only signals, BSGA dynamically gates soft reward priors to modulate CoM height and lower-limb coordination based on the estimated slope geometry -- encouraging hip-dominant uphill propulsion and knee-oriented downhill braking. Crucially, the deployed actor remains entirely proprioceptive, requiring no online exteroceptive sensing. Extensive Sim-to-Real experiments demonstrate that our framework effectively mitigates posture degeneration and enables blind, continuous traversal of outdoor grass slopes up to 62.7% ($32.1^\circ$), validating a physics-guided approach to challenging slope terrain adaptation.
