Haodong Zhang

I am currently a Researcher at Tencent Robotics X, focusing on humanoid robot whole-body control and dexterous manipulation. I received my Ph.D. degree from Zhejiang University in 2025, under the supervision of Professor Rong Xiong and Yue Wang, where my research focused on motion imitation learning. Prior to that, I received my B.S. degree from Chu Kochen Honors College (Mixed Class), Zhejiang University, Hangzhou, China, in 2020, with a major in Control Science and Engineering.

Email / Scholar / Github

Research

My latest research interests include embodied AI and imitation learning.

* denotes equal contribution, † denotes corresponding author.

	Natural Humanoid Robot Locomotion with Generative Motion Prior Haodong Zhang, Liang Zhang, Zhenghan Chen, Lu Chen, Yue Wang, Rong Xiong† IEEE International Conference on Intelligent Robots and Systems (IROS), 2025 project page/ arXiv The generative motion prior (GMP) provides fine-grained motion-level supervision for the task of natural humanoid robot locomotion.
	EMP: Executable Motion Prior for Humanoid Robot Standing Upper-body Motion Imitation Haocheng Xu, Haodong Zhang, Zhenghan Chen, Rong Xiong† arXiv preprint arXiv:2507.15649, 2025 arXiv An executable motion prior (EMP) enables humanoid robots to imitate diverse standing upper-body motions while preserving physical feasibility.
	A Whole-Body Motion Imitation Framework from Human Data for Full-Size Humanoid Robot Zhenghan Chen, Haodong Zhang, Dong Wang, Jiayi Yu, Haocheng Xu, Yue Wang, Rong Xiong† IEEE International Conference on Real-time Computing and Robotics (RCAR), 2025 arXiv A whole-body motion imitation framework that transfers human motion data to a full-size humanoid robot.
	Semantics-aware motion retargeting with vision-language models Haodong Zhang, Zhike Chen, Haocheng Xu, Lei Hao, Xiaofei Wu, Songcen Xu, Zhensong Zhang, Yue Wang, Rong Xiong† Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2024 project page/ arXiv/ code We introduce the incorporation of robust, state-of-the-art vision-language models to provide semantic guidance to the motion retargeting network.
	Learning active force–torque based policy for sub-mm localization of unseen holes Liang Xie, Haodong Zhang, Yinghao Zhao, Yu Zhang, Rong Xiong, Yi Ren, Yue Wang† IEEE Transactions on Industrial Informatics, 2024 paper/ code Our goal is to enable robots to operate effectively in contact-rich environments with tight tolerances, and adapt to new tasks involving unseen peg-hole pairs.
	Unified Cross-Structural Motion Retargeting for Humanoid Characters Haodong Zhang, Zhike Chen, Haocheng Xu, Lei Hao, Xiaofei Wu, Songcen Xu, Rong Xiong†, Yue Wang† IEEE Transactions on Visualization and Computer Graphics (TVCG), 2024 paper We address the challenge of motion retargeting across previously unseen skeletal structures with a unified dynamic graph network.
	Human-Robot Sign Language Motion Retargeting from Videos Haodong Zhang, Haocheng Xu, Yue Wang, Rong Xiong† 43rd Chinese Control Conference (CCC), 2024 paper A pipeline to retarget human sign language motions from videos onto a humanoid robot for expressive sign language communication.
	Robust Real-Time Motion Retargeting via Neural Latent Prediction Tiantian Wang, Haodong Zhang, Lu Chen, Dongqi Wang, Yue Wang, Rong Xiong† IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2023 paper Real-time motion retargeting via neural latent prediction for robust human-to-robot motion transfer.
	DPCN++: Differentiable Phase Correlation Network for Versatile Pose Registration Zexi Chen, Yiyi Liao, Haozhe Du, Haodong Zhang, Xuecheng Xu, Haojian Lu, Rong Xiong, Yue Wang† IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), 2023 paper/ arXiv A differentiable phase correlation network that enables versatile pose registration across 2D and 3D modalities.
	Learning to Fill the Seam by Vision: Sub-millimeter Peg-in-hole on Unseen Shapes in Real World Liang Xie, Hongxiang Yu, Yinghao Zhao, Haodong Zhang, Zhongxiang Zhou, Minhang Wang, Yue Wang, Rong Xiong† IEEE International Conference on Robotics and Automation (ICRA), 2022 arXiv/ code By imitating the human's behavior, we design architectures with position and orientation estimators based on the seam representation for pose alignment, which proves to be general to the unseen peg geometries.
	Kinematic motion retargeting via neural latent optimization for learning sign language Haodong Zhang, Weijie Li, Jiangpin Liu, Zexi Chen, Yue Wang†, Rong Xiong† IEEE Robotics and Automation Letters (RAL), 2022 presented at IEEE International Conference on Robotics and Automation (ICRA), 2022 arXiv/ code Latent optimization utilizes a decoder to establish a mapping between the latent space and the robot motion space. Afterward, the retargeting results that satisfy robot constraints can be obtained by searching for the optimal latent vector.
	Learning world transition model for socially aware robot navigation Yuxiang Cui, Haodong Zhang, Yue Wang, Rong Xiong† IEEE International Conference on Robotics and Automation (ICRA), 2021 arXiv/ code A model-based reinforcement learning approach for robots to navigate through crowded environments.
	Champion Team Paper: Dynamic Passing-Shooting Algorithm of the RoboCup Soccer SSL 2019 Champion Zexi Chen, Haodong Zhang, Dashun Guo, Shenhan Jia, Xianze Fang, Zheyuan Huang, Yunkai Wang, Peng Hu, Licheng Wen, Lingyun Chen, Zhengxi Li, Rong Xiong† RoboCup 2019: Robot World Cup XXIII, 2019 paper/ arXiv The dynamic passing-shooting algorithm developed by ZJUNlict, champion of the RoboCup Soccer Small Size League (SSL) 2019.