Research Assistant Professor
University of Michigan Transportation Research Institute (UMTRI)
I am a Research Assistant Professor at the University of Michigan Transportation Research Institute (UMTRI).
Before joining UMTRI, I was a Lecturer in the Department of Civil and Environmental Engineering at the University of Michigan, Ann Arbor. I was also a Postdoctoral Research Fellow in the Mobility Transformation Lab, working with Prof. Henry X. Liu.
I received my Ph.D. and B.E. from Tsinghua University in 2023 and 2018, respectively, with my doctoral research advised by Prof. Keqiang Li. From Dec 2022 to Dec 2023, I was a visiting PhD student in the Automatic Control Laboratory at EPFL (École Polytechnique Fédérale de Lausanne), advised by Prof. Colin Jones. During my doctoral research, I also received guidance from Prof. Yang Zheng at UC San Diego.
My research lies at the intersection of machine learning, control, optimization, and simulation, addressing how autonomous systems can operate safely and efficiently alongside people in complex urban environments. With a primary focus on autonomous driving, I develop generative AI methods for high-fidelity simulation and rigorous safety evaluation, as well as reliable and scalable data-driven control strategies. Through these efforts, I aim to advance safe, efficient, and sustainable urban mobility.
† equal contribution · * corresponding author · See Google Scholar for the full list.
An automated pipeline for synthesizing safety-critical data for safe and scalable deployment of end-to-end autonomous driving at anywhere in the world.
A simulation framework that generates statistically realistic and risk-adjustable traffic scenes using multi-agent conditional diffusion for stress testing of autonomous vehicle safety.
An open-source, high-fidelity traffic simulation platform designed to uncover unknown unsafe events and efficiently estimate AV statistical performance metrics.
A decentralized robust data-driven predictive control framework for CAVs to smooth mixed traffic flow while ensuring computational scalability.
Robust data-driven predictive control framework for mixed platoons using reachability analysis to handle noise and attacks while ensuring safety.
First experimental validation of data-driven predictive control for CAVs in dissipating traffic waves using miniature experiment platform.
A cooperative distributed data-driven predictive control framework for CAVs in large-scale mixed traffic flow.
A data-driven nonparametric strategy for safe and optimal control of CAVs in mixed traffic using Willems' fundamental lemma and receding horizon optimization.
A miniature experimental platform MCCT based on Mixed Digital Twin concept for validating multi-vehicle cooperation and vehicle-road-cloud integration.
Investigation on how the information flow topology ("looking ahead" or "looking behind") and the maximum platoon size influence the stability of mixed traffic flow.
A novel Leading Cruise Control (LCC) framework for CAVs to actively lead the motion of the vehicles behind, while maintaining the car-following operations to the vehicles ahead.
Investigation of CAV formation patterns in mixed traffic from set-function optimization perspective, revealing optimal formations beyond platooning for system-level traffic benefits.
First rigorous proof of controllability and stabilizability properties of mixed traffic systems via CAVs with heterogeneous human-driven vehicles.
Proposes "1+n" control framework for CAV control at signalized intersections in mixed traffic, enabling the CAVs to significantly improve the traffic energy efficiency at a low penetration rate.
First rigorous theoretical analysis of controllability, stabilizability, and reachability of mixed traffic systems, showing that CAVs can effectively improve traffic with only 5% penetration rate.
