Legged Robots / Behavior Foundation Models / World Models

Zifan Wang

PhD student in Artificial Intelligence at The Hong Kong University of Science and Technology (Guangzhou), supervised by Prof. Junwei Liang. I work on legged robots, including quadrupeds, humanoids, and wheeled-legged robots, with a focus on optimal control, reinforcement learning, loco-manipulation, and robot-centric world modeling.

Currently on the job market. Open to roles in visual behavior foundation models and humanoid world models.

Research Legged robots, locomotion, reinforcement learning, loco-manipulation

Venues RSS, NeurIPS, CoRL, CVPR, IROS

Stack Python, C/C++, ROS/ROS2, STM32, Jetson, SolidWorks, Altium Designer

Service Reviewer for T-RO, RAL, IROS, CoRL, ICCV, Humanoids

News

2026.09: 🎉 Four papers accepted by CoRL 2026: Perceptive Behavior Foundation Model, UniLab, MeshMimic, and SOLO.
2026: 🎉 Two papers accepted by the RSS 2026 Workshop on Whole-body Control and Bimanual Manipulation (Perceptive BFM & UniLab).
2026: New preprint: Perceptive Behavior Foundation Model (arXiv, Code) — adapting human motion priors to robot-centric terrain.
2026: One paper accepted by RSS 2026: GS-Playground (Code), a high-throughput photorealistic simulator for vision-informed robot learning.
2026: New preprint: UniLab (Code) — a heterogeneous architecture for robot RL beyond GPU-dominant paradigms.
2026: New preprint: DiT4DiT (Code) — jointly modeling video dynamics and actions for generalizable robot control.
2025: One paper accepted by NeurIPS 2025 on vision-language-action cross-task generalization.
2025: Two papers accepted by CoRL 2025 (OmniPerception & GLOVER++).
2025: Two papers accepted by IROS 2025 (Diffusion Planner & DISCOVERSE).
2025: One paper accepted by CVPR 2025 on mitigating human-robot domain discrepancy.
2024: One paper accepted by CoRL 2024 on contrastive imitation learning.
2024: One paper accepted by IROS 2024 on loco-manipulation for wheel-legged robots.

GitHub Stats

-- Total Stars

-- Total Forks

6 Tracked Repos

PMT

Research Code

Official implementation of Perceptive Behavior Foundation Model.

-- --

Project Page Code

UniLab

Research Code

Heterogeneous architecture for robot RL beyond GPU-dominant paradigms.

-- --

Project Page Code

GS-Playground

Research Code

High-throughput photorealistic simulator for vision-informed robot learning.

-- --

Project Page Code

DiT4DiT

Research Code

Joint video dynamics and action modeling for generalizable robot control.

-- --

Project Page Code

Awesome Loco-Manipulation

Open Source

Curated resources for loco-manipulation research and engineering.

-- --

Code

Genesis Legged Gym

Open Source

Legged robot training framework for the Genesis simulator.

-- --

Code

Publications

Locomotion

Perceptive Behavior Foundation Model: Adapting Human Motion Priors to Robot-Centric Terrain

Z Wang, Y Li, T Ma, Q Zhang, Y Fan, H Xu, S Yang, J Liang

CoRL 2026 RSS 2026 WCBM Workshop

Project Page arXiv Code

Omni-Perception: Omnidirectional Collision Avoidance for Legged Locomotion in Dynamic Environments

Z Wang, T Ma, Y Jia, X Yang, J Zhou, W Ouyang, Q Zhang, J Liang

CoRL 2025 Oral

Project Page Code

End-to-End Humanoid Robot Safe and Comfortable Locomotion Policy

Z Wang, X Yang, J Zhao, J Zhou, T Ma, Z Gao, A Ajoudani, J Liang

arXiv 2025

Preference Aligned Diffusion Planner for Quadrupedal Locomotion Control

X Yuan, Z Shang, Z Wang, C Wang, Z Shan, M Zhu, C Bai, X Li, W Wan, et al.

IROS 2025

Project Page

MeshMimic: Geometry-Aware Humanoid Motion Learning through 3D Scene Reconstruction

Q Zhang, J Ma, P Liu, S Shi, Z Su, Z Wang, J Sun, W Cui, J Yu, G Han, et al.

CoRL 2026

SOLO: Stable Omni-terrain Long-Horizon Perceptive Humanoid Locomotion

P Sun, G Han, J Sun, J Ma, Z Su, Z Tao, P Liu, S Shi, W Cui, Z Wang, J Yu, W Zhao, et al.

CoRL 2026

Loco-Manipulation

Arm-Constrained Curriculum Learning for Loco-Manipulation of a Wheel-Legged Robot

Z Wang, Y Jia, L Shi, H Wang, H Zhao, X Li, J Zhou, J Ma, G Zhou

IROS 2024 Oral

Code

Manipulation & Imitation Learning

DiT4DiT: Jointly Modeling Video Dynamics and Actions for Generalizable Robot Control

T Ma, J Zheng, Z Wang, C Jiang, A Cui, J Liang, S Yang

arXiv 2026

Project Page arXiv Code

Exploring the Limits of Vision-Language-Action Manipulations in Cross-Task Generalization

J Zhou, K Ye, J Liu, T Ma, Z Wang, R Qiu, KY Lin, Z Zhao, J Liang

NeurIPS 2025

GLOVER++: Unleashing the Potential of Affordance Learning from Human Behaviors for Robotic Manipulation

T Ma, J Zheng, Z Wang, Z Gao, J Zhou, J Liang

CoRL 2025

Project Page

Contrastive Imitation Learning for Language-guided Multi-Task Robotic Manipulation

T Ma, J Zhou, Z Wang, R Qiu, J Liang

CoRL 2024

Project Page

Mitigating the Human-Robot Domain Discrepancy in Visual Pre-training for Robotic Manipulation

J Zhou, T Ma, KY Lin, Z Wang, R Qiu, J Liang

CVPR 2025

Project Page

GLOVER: Generalizable Open-Vocabulary Affordance Reasoning for Task-Oriented Grasping

T Ma, Z Wang, J Zhou, M Wang, J Liang

CoRL 2024 GenPriors Workshop Best Paper Award

Project Page

Simulation

GS-Playground: A High-Throughput Photorealistic Simulator for Vision-Informed Robot Learning

Y Jia, H Zhang, Z Zhang, J Wu, M Yu, Z Wang, et al.

RSS 2026

Project Page arXiv Code

UniLab: A Heterogeneous Architecture for Robot RL Beyond GPU-Dominant Paradigms

Y Jia, Z Cao, M Yu, H Zhang, S Chen, D Jiang, ..., Z Wang, et al.

CoRL 2026 RSS 2026 WCBM Workshop

Project Page arXiv Code

DISCOVERSE: Efficient Robot Simulation in Complex High-Fidelity Environments

Y Jia, G Wang, Y Dong, J Wu, Y Zeng, H Lin, Z Wang, H Ge, W Gu, K Ding, et al.

IROS 2025 Oral

Project Page

FGGS-LiDAR: Ultra-Fast, GPU-Accelerated Simulation from General 3DGS Models to LiDAR

J Wu, Y Jia, Y Yan, Z Chen, T Tan, Z Wang, G Wang

arXiv

Open Source Projects

Awesome Loco-Manipulation

Zifan Wang

Open Source

Curated resources for loco-manipulation research and engineering.

Code

Genesis Legged Gym

Zifan Wang

Open Source

Legged robot training framework for the Genesis simulator.

Code

Education

     PhD in Artificial Intelligence
    HKUST (Guangzhou)
2024 - 2027 (Expected)
Supervisor: Junwei Liang

  

     MSc in Robotics & Autonomous Systems
    HKUST (Guangzhou)
2022 - 2024, GPA: 3.84
Supervisors: Ming Liu, Jun Ma

  

     BEng in Automation
    Southwest Jiaotong University
2017 - 2021, Score: 85.45
School of Information Science and Technology

  

Honors & Awards

First Class Scholarship — Southwest Jiaotong University
The Enterprise Scholarship — Southwest Jiaotong University
Outstanding Students Cadre — Southwest Jiaotong University

Competitions

National Second Prize — ABU Robocon (Equestrian Simulation)

Asia-Pacific Broadcasting Union Robocon, 2020

Developed Simulink simulation system and co-simulation with Adams for quadruped robot maneuvers.

National Third Prize — RoboMaster (DJI)

RoboMaster, 2020

Developed LQR-based control framework, image processing and LiDAR sensing algorithms, sensor fusion, and hardware circuit design.

National Second Prize — ABU Robocon

Asia-Pacific Broadcasting Union Robocon, 2019

Responsible for quadruped robot leg motion control and upper structure ejection/gripping device control.

Provincial Second Prize — Engineering Training Competition

National College Student Engineering Training Competition, 2019

Hardware circuit design, sensor data fusion with orthogonal code disk and DT35 laser ranging module.

Provincial Third Prize — Electronics Design Contest

National Undergraduate Electronics Design Contest, 2019

Electromagnetic gun hardware circuit design, BP neural network-based control algorithm, and UCOS-III RTOS porting on ARM processor.

Contact: wang_zifan@outlook.com / zwang765@connect.hkust-gz.edu.cn / +86 135 1839 3890