Learn to Teach: Sample-Efficient Privileged Learning for Humanoid Locomotion Over Real-World Uneven Terrain
- 23 July 2025
- journal article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Robotics and Automation Letters
- Vol. 10 (9) , 9048-9055
Abstract
Humanoid robots promise transformative capabilities for industrial and service applications. While recent advances in Reinforcement Learning (RL) yield impressive results in locomotion, manipulation, and navigation, the proposed methods typically require enormous simulation samples to account for real-world variability. This work proposes a novel one-stage training framework—Learn to Teach (L2T)—which unifies teacher and student policy learning. Our approach recycles simulator samples and synchronizes the learning trajectories through shared dynamics, significantly reducing sample complexities and training time while achieving state-of-the-art performance. Furthermore, we validate the RL variant (L2T-RL) through extensive simulations and hardware tests on the Digit robot, demonstrating zero-shot sim-to-real transfer and robust performance over 12+ diverse terrains without depth estimation modules.Keywords
All Related Versions
Funding Information
- Office of Naval Research (#0023 N000142312223)
- National Science Foundation (#0023 IIS-1924978, #0023 CMMI-2144309, #0023 FRR-2328254)
- U.S. Department of Agriculture (2023-67021-41397)
This publication has 20 references indexed in Scilit:
- CTS: Concurrent Teacher-Student Reinforcement Learning for Legged LocomotionIEEE Robotics and Automation Letters, 2024
- Open X-Embodiment: Robotic Learning Datasets and RT-X Models : Open X-Embodiment Collaboration02024
- Infer and Adapt: Bipedal Locomotion Reward Learning from Demonstrations via Inverse Reinforcement Learning2024
- Sim-to-Real Learning for Humanoid Box Loco-Manipulation2024
- Learning Arm-Assisted Fall Damage Reduction and Recovery for Legged Mobile Manipulators2023
- Orbit: A Unified Simulation Framework for Interactive Robot Learning EnvironmentsIEEE Robotics and Automation Letters, 2023
- Policy mirror descent for reinforcement learning: linear convergence, new sampling complexity, and generalized problem classesMathematical Programming, 2022
- Learning quadrupedal locomotion over challenging terrainScience Robotics, 2020
- Sim-to-Real Transfer of Robotic Control with Dynamics Randomization2018
- Self-improving reactive agents based on reinforcement learning, planning and teachingMachine Learning, 1992