Zilin Huang
Research Associate
Department of Civil and Environmental Engineering
University of Wisconsin-Madison
Office: Engineering Hall 1217
Email: zilin.huang@wisc.edu
CV: Hi, thanks for stopping by! 
I am currently a Research Associate at the 
University of Wisconsin–Madison and a Visiting Research Fellow at 
Purdue University.
I am a founding member of 
Sky-Lab and a member of the 
Purdue Autonomous Go-Kart Initiative (PAGKI).
I am also affiliated with the 
Center for Connected and Automated Transportation (CCAT), the 
Smart Highway Research Center (SHRC), and the 
Tribal and Rural Autonomous Vehicles for Efficiency, Livability and Safety (TRAVELS) Center.
I received my Ph.D. and M.S. degrees from the University of Wisconsin–Madison, focusing on foundation models and human-guided reinforcement learning for autonomous driving, under the supervision of Prof. Sikai (Sky) Chen. Before joining UW–Madison, I worked as a Research Assistant with Prof. Samuel Labi at Purdue University. My work is driven by an interest in translating cutting-edge research into practical, real-world applications. Prior to my Ph.D. studies, I founded two technology companies in China.
📢 I am on the 2026–2027 academic job market and seeking faculty positions.
Research Statement
My research lies at the intersection of Physical AI and human-centered transportation systems. I envision future mobility as a human-robot-society ecosystem, where humans coexist, interact, and collaborate with diverse autonomous agents (e.g., autonomous vehicles, delivery robots, drones, and flying vehicles). These agents must understand complex physical and social environments, reason about their actions, adapt to human preferences and social norms, and remain safe and reliable under open-world uncertainty.
Building on this vision, my research aims to develop Human-centered and trustworthy autonomous (HEART) systems that can safely operate, interact, and continuously improve alongside humans in the physical world. This agenda centers on three questions: 1) How can autonomous agents understand complex physical and social environments and make safe, reliable, and socially aware decisions? 2) How can capabilities learned in simulation or from large-scale pretrained models be transferred and grounded in real physical systems across differences in embodiment, sensing, and hardware? 3) How can autonomous agents achieve safe recursive self-improvement through physical interaction, failure discovery, and real-world feedback?
To address these questions, I integrate VLMs, VLAs, and world models with reinforcement learning, cross-embodiment learning, Sim2Real transfer, closed-loop simulation, and real-world deployment. Ultimately, my goal is to advance safer, more efficient, and human-centered transportation systems where humans and autonomous agents can coexist and collaborate in the physical world.
Recent News View All
| Mar 20, 2025 |
|
🎉 We received Two NVIDIA Academic Grant Program Awards ($43,800 USD).
We received two NVIDIA grants for AI research in transportation - 20,000 A100 GPU hours for SafetyGPT and two RTX PRO 6000 Blackwell GPUs for EdgeTwin Project. Read more →
|
|---|---|---|
| Jan 12, 2025 |
|
👏 I attended the TRB 2025 Annual Meeting.
I participated in the Transportation Research Board (TRB) 104th Annual Meeting (January 7-11) at the Washington, D.C. Convention Center. We presented our work in multiple poster and lectern sessions, engaging... Read more →
|
| Jan 10, 2025 |
|
👏 I was invited to visit Google
I was invited to visit Google’s office in Washington DC on January 10, 2025. During the visit, I had the opportunity to meet with Google’s research team and discuss potential... Read more →
|
| Oct 18, 2024 |
|
🎉 Paper Traffic expertise meets residual RL: Knowledge-informed model-based residual reinforcement learning for CAV trajectory control is published on Communications in Transportation Research. |
| Sep 1, 2024 |
|
📢 We currently released one paper: Trustworthy Human-AI Collaboration: Reinforcement Learning with Human Feedback and Physics Knowledge for Safe Autonomous Driving. |
View All
- Preprint
Sim2Real-AD: A Modular Sim-to-Real Framework for Deploying VLM-Guided Reinforcement Learning in Real-World Autonomous Driving 🔥arXiv preprint arXiv:2604.03497 (Preprint) , 2026 - Preprint
DriveVLM-RL: Neuroscience-Inspired Reinforcement Learning with Vision-Language Models for Safe and Deployable Autonomous Driving 🔥arXiv preprint arXiv:2603.18315 (Preprint) , 2026
