Currently studying and researching at Shanghai Jiao Tong University, with a steady focus on embodied intelligence, VLA, robotics systems, and reinforcement learning.
Research Experience
IWIN-FINS Lab, Shanghai Jiao Tong University
Feb 2025 - Present
Undergraduate Researcher - Underwater Robot Learning | Advisor: Prof. Jianping He
- Lead the end-to-end development of FinsSim, a modular simulation-to-hardware framework for learning-based underwater robot control centered on the FinsROV, spanning simulation, learning, perception and state estimation, actuation calibration, and ROS 2 deployment.
- Built Unity/MARUS and Isaac Lab simulation paths with interchangeable hydrodynamic models, domain randomization, and reusable RL/MARL interfaces; implemented PPO-based control with Wrench6 and Thruster8 action spaces for station keeping and 3D trajectory tracking.
- Developed the pool-scale sim-to-real stack, including refractive AprilTag localization, EKF state fusion, calibrated thruster/motor and hydrodynamic models, constrained 6x8 wrench allocation, and a ROS 2 hardware safety bridge; conducted matched simulation and real-pool experiments.
- Selected for the 2026 Zhiyuan Future Scholars Program with an independent undergraduate research project on underwater reinforcement learning.
RHOS Lab, Shanghai Jiao Tong University
Sep 2025 - Present
Undergraduate Researcher - VLA Pretraining & Robot Learning | Advisor: Prof. Yong-Lu Li
- Contributed to GM-100, a 100-task real-world benchmark for embodied AI agents, with a focus on robot trajectory and model-data collection and baseline VLA model training; fifth author on the public preprint.
- Conduct follow-up VLA pretraining work using human-hand keypoint datasets, including data processing and training keypoint-extraction models for the pretraining pipeline.
Publications
- Yu Zhang, Yuanmingqing Song, et al. “FinsSim: A reality-aligned integrated simulation platform for underwater robot learning.” Manuscript under review at IEEE International Conference on Robotics and Automation (ICRA). [First author]
- Mingwei Li, Dehao Ye, Yu Zhang, Ziyu Wang, Yong-Lu Li. “Data That Selects Itself: Scalable Self-Curation for Robot Learning.” Manuscript under review at International Conference on Learning Representations (ICLR 2027), 2026. [Third author]
- Ziyu Wang, Chenyuan Liu, Yushun Xiang, Runhao Zhang, Yu Zhang, et al. “The Great March 100: 100 Detail-oriented Tasks for Evaluating Embodied AI Agents.” arXiv:2601.11421, 2026. [Fifth author; manuscript under review]