Gait-Adaptive Perceptive Humanoid Locomotion with Real-Time Under-Base Terrain Reconstruction

2025-12-08

AI Summary

To address rapid balance loss in full-scale humanoid robots navigating complex terrains—such as long staircases and large gaps—caused by ambiguous perception, gait timing misalignment, and insufficient whole-body coordination, this paper proposes an end-to-end reinforcement learning framework integrating real-time terrain perception and adaptive gait regulation. Methodologically, we employ a U-Net architecture to reconstruct dense bottom-depth maps for precise height estimation, fusing downward visual input with proprioceptive sensor data; gait timing and whole-body motion policies are jointly optimized within a single-stage continuous teacher–student training paradigm. Our key contribution is a unified perception–control policy architecture enabling terrain-driven, concurrent adaptation of gait and posture. Experimental validation on a 31-DOF, 1.65 m tall humanoid demonstrates successful execution of challenging tasks—including ascending/descending long staircases and backward traversal over a 46 cm gap—significantly enhancing dynamic balance robustness across diverse indoor and outdoor terrains.

Technology Category

• Intelligent Robots: Behavior Learning & Control
• Humans and AI: Human-Aware Planning and Behavior Prediction
• Planning, Routing, and Scheduling: Learning for Planning and Scheduling

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

For full-size humanoid robots, even with recent advances in reinforcement learning-based control, achieving reliable locomotion on complex terrains, such as long staircases, remains challenging. In such settings, limited perception, ambiguous terrain cues, and insufficient adaptation of gait timing can cause even a single misplaced or mistimed step to result in rapid loss of balance. We introduce a perceptive locomotion framework that merges terrain sensing, gait regulation, and whole-body control into a single reinforcement learning policy. A downward-facing depth camera mounted under the base observes the support region around the feet, and a compact U-Net reconstructs a dense egocentric height map from each frame in real time, operating at the same frequency as the control loop. The perceptual height map, together with proprioceptive observations, is processed by a unified policy that produces joint commands and a global stepping-phase signal, allowing gait timing and whole-body posture to be adapted jointly to the commanded motion and local terrain geometry. We further adopt a single-stage successive teacher-student training scheme for efficient policy learning and knowledge transfer. Experiments conducted on a 31-DoF, 1.65 m humanoid robot demonstrate robust locomotion in both simulation and real-world settings, including forward and backward stair ascent and descent, as well as crossing a 46 cm gap. Project Page:https://ga-phl.github.io/

Problem

• Achieving reliable humanoid locomotion on complex terrains
• Integrating terrain sensing, gait regulation, and whole-body control
• Adapting gait timing and posture to terrain geometry in real-time

Innovation

• Real-time dense height map reconstruction via U-Net
• Unified policy for gait timing and whole-body posture adaptation
• Single-stage successive teacher-student training scheme