COLA: Learning Human-Humanoid Coordination for Collaborative Object Carrying
Yushi Du, Yixuan Li, Baoxiong Jia, Yutang Lin, Pei Zhou, Wei Liang, Yanchao Yang, Siyuan Huang
AI summary
Problem
Enabling compliant human-humanoid collaboration for object carrying remains largely unexplored due to complex whole-body dynamics, limited force sensing, and the challenge of dynamically coordinating leader-follower roles in real-time.
Approach
COLA trains a residual teacher policy with privileged object-state information in a closed-loop simulator, then distills it into a student policy that relies solely on the robot's proprioception to infer interaction forces and coordinate movement.
Key results
- Reduces human effort by 24.7% compared to baselines in simulation
- Achieves 10.2 cm/s linear and 0.1 rad/s angular tracking error relative to human motion
- Validates robust real-world carrying across diverse objects and terrains without external sensors
- Human user studies confirm 27.4% improvement in collaboration compliance
Why it matters
Offers a practical, sensor-free framework for deploying compliant humanoid robots in real-world collaborative tasks like healthcare, domestic assistance, and logistics.
Abstract
Human-humanoid collaboration shows significant promise for applications in healthcare, domestic assistance, and manufacturing. While compliant robot-human collaboration has been extensively developed for robotic arms, enabling compliant human-humanoid collaboration remains largely un- explored due to humanoids’ complex whole-body dynamics. In this paper, we propose a proprioception-only reinforcement learning approach, COLA, that combines leader and follower behaviors within a single policy. The model is trained in a closed-loop environment with dynamic object interactions to ∗Equal contributions. † Corresponding authors. Emails: jiabaoxiong@bigai.ai, liangwei@bit.edu.cn, yanchaoy@hku.hk, and syhuang@bigai.ai 1 The University of Hong Kong 2 State Key Laboratory of General Artifi- cial Intelligence, Beijing Institute for General Artificial Intelligence (BIGAI) 3 Beijing Institute of Technology 4 Yuanpei College, Peking University 5 Yangtze Delta Region Academy of Beijing Institute of Technology This work is supported by the Early Career Scheme of the Research Grants Council (grant # 27207224), the National Natural Science Foundation of China (NSFC) under Grant No.62172043, and in part by the JC STEM Lab of Autonomous Intelligent Systems funded by The Hong Kong Jockey Club Charities Trust. predict object motion patterns and human intentions implicitly, enabling compliant collaboration to maintain load balance through coordinated trajectory planning. We evaluate our approach through comprehensive simulator and real-world experiments on collaborative carrying tasks, demonstrating the effectiveness, generalization, and robustness of our model across various terrains and objects. Simulation experiments demonstrate that our model reduces human effort by 24.7% compared to baseline approaches while maintaining object stability. Real-world experiments validate robust collaborative carrying across different object types (boxes, desks, stretch- ers, etc.) and movement patterns (straight-line, turning, slope climbing). Human user studies with 23 participants confirm an average improvement of 27.4% compared to baselines. Our method enables compliant human-humanoid collaborative car- rying without requiring external sensors or complex interaction models, offering a practical solution for real-world deployment.