TactileStep: Sole Tactile Learning for Regulating Foot-Terrain Interaction in Humanoid Locomotion

2026-09-23

AI Summary

This study addresses the substantial landing impacts and unstable contacts experienced by humanoid robots navigating complex terrains due to the absence of plantar tactile feedback. To this end, this work introduces plantar tactile sensing into locomotion control for the first time, proposing a deployable tactile learning framework. By aligning features between simulated and real-world pressure insoles, the method integrates tactile and proprioceptive cues to identify contact phases, while designing phase-aware rewards to optimize reinforcement learning policies. Experimental evaluations on the Unitree G1 humanoid robot demonstrate that the proposed approach reduces peak landing forces by 48.8%, decreases impact noise by 30.1 dB, and increases the standing contact area by 23.8%. These results indicate significant improvements in both the safety and stability of robotic locomotion.

Abstract

Humanoid parkour policies can traverse various terrains, but task completion may mask challenges of harsh landings, edge contacts, and unstable stance contacts. Humans naturally regulate foot-terrain interaction through tactile feedback, modulating contact compliance according to terrain stiffness. This highlights a key domain gap between humans and humanoid robots: the absence of rich tactile sensing in most humanoid systems. We address this problem with TactileStep, a deployable tactile learning framework that brings sole pressure sensing into humanoid locomotion control for softer touchdowns and more stable support. TactileStep aligns tactile simulation with the real pressure insole, allowing the policy to learn from the same contact features available on hardware. During training, we use tactile and motion cues to recognize different foot-contact phases and apply phase-aware rewards that encourage safer landing and more stable stance. Evaluated in simulation and on a Unitree G1 humanoid across diverse terrains, TactileStep reduces peak touchdown force by up to 48.8% and peak A-weighted impact noise by up to 30.1 dB over a strong perceptive baseline, while increasing stance contact area by up to 23.8%.

Problem

• Humanoid Locomotion
• Foot-Terrain Interaction
• Tactile Sensing
• Sole Pressure
• Contact Stability

Innovation

• Tactile Learning
• Humanoid Locomotion
• Sole Pressure Sensing
• Sim-to-Real Alignment
• Phase-Aware Rewards