Learning Humanoid Locomotion with World Model Reconstruction

TL;DR

This paper introduces World Model Reconstruction (WMR), an end-to-end learning approach enabling humanoid robots to navigate complex terrains blindly by reconstructing the environment and improving locomotion robustness.

Contribution

The study presents a novel joint training method for world state reconstruction and locomotion policy, enhancing real-world terrain navigation without relying on sensor data.

Findings

Successfully navigated 3.2 km across icy and snowy terrains

Demonstrated robustness on deformable and slippery surfaces

Achieved autonomous long-distance hiking in real-world conditions

Abstract

Humanoid robots are designed to navigate environments accessible to humans using their legs. However, classical research has primarily focused on controlled laboratory settings, resulting in a gap in developing controllers for navigating complex real-world terrains. This challenge mainly arises from the limitations and noise in sensor data, which hinder the robot's understanding of itself and the environment. In this study, we introduce World Model Reconstruction (WMR), an end-to-end learning-based approach for blind humanoid locomotion across challenging terrains. We propose training an estimator to explicitly reconstruct the world state and utilize it to enhance the locomotion policy. The locomotion policy takes inputs entirely from the reconstructed information. The policy and the estimator are trained jointly; however, the gradient between them is intentionally cut off. This ensures that the estimator focuses solely on world reconstruction, independent of the locomotion policy's updates. We evaluated our model on rough, deformable, and slippery surfaces in real-world scenarios, demonstrating robust adaptability and resistance to interference. The robot successfully completed a 3.2 km hike without any human assistance, mastering terrains covered with ice and snow.