Towards Adaptive Humanoid Control via Multi-Behavior Distillation and Reinforced Fine-Tuning (original) (raw)

Authors

• Yingnan Zhao College of Computer Science and Technology, Harbin Engineering University National Engineering Laboratory for Modeling and Emulation in E-Government, Harbin Engineering University
• Xinmiao Wang College of Computer Science and Technology, Harbin Engineering University Institute of Artificial Intelligence (TeleAI), China Telecom
• Dewei Wang Institute of Artificial Intelligence (TeleAI), China Telecom School of Information Science and Technology, University of Science and Technology of China
• Xinzhe Liu Institute of Artificial Intelligence (TeleAI), China Telecom School of Information Science and Technology, ShanghaiTech University
• Dan Lu College of Computer Science and Technology, Harbin Engineering University National Engineering Laboratory for Modeling and Emulation in E-Government, Harbin Engineering University
• Qilong Han College of Computer Science and Technology, Harbin Engineering University National Engineering Laboratory for Modeling and Emulation in E-Government, Harbin Engineering University
• Peng Liu College of Computer Science and Technology, Harbin Institute of Technology
• Chenjia Bai Institute of Artificial Intelligence (TeleAI), China Telecom Shenzhen Research Institute of Northwestern Polytechnical University

DOI:

https://doi.org/10.1609/aaai.v40i22.38951

Abstract

Humanoid robots are promising to learn a diverse set of human-like locomotion behaviors, including standing up, walking, running, and jumping. However, existing methods predominantly require training independent policies for each skill, yielding behavior-specific controllers that exhibit limited generalization and brittle performance when deployed on irregular terrains and in diverse situations. To address this challenge, we propose Adaptive Humanoid Control (AHC) that adopts a two-stage framework to learn an adaptive humanoid locomotion controller across different skills and terrains. Specifically, we first train several primary locomotion policies and perform a multi-behavior distillation process to obtain a basic multi-behavior controller, facilitating adaptive behavior switching based on the environment. Then, we perform reinforced fine-tuning by collecting online feedback in performing adaptive behaviors on more diverse terrains, enhancing terrain adaptability for the adaptive behavior controller. We conduct experiments in both simulation and real-world experiments in Unitree G1 robots. The results show that our method exhibits strong adaptability across various situations and terrains.

How to Cite

Zhao, Y., Wang, X., Wang, D., Liu, X., Lu, D., Han, Q., Liu, P., & Bai, C. (2026). Towards Adaptive Humanoid Control via Multi-Behavior Distillation and Reinforced Fine-Tuning. Proceedings of the AAAI Conference on Artificial Intelligence, 40(22), 18818-18826. https://doi.org/10.1609/aaai.v40i22.38951

Issue

Section

AAAI Technical Track on Intelligent Robotics