Research

I embed traffic-flow theory into machine learning to model and estimate traffic dynamics at the network scale. This work includes physics-informed formulations of macroscopic models such as the generalized bathtub model for large-scale urban networks, graph learning guided by the network macroscopic fundamental diagram for traffic state imputation, and observation-aligned two-stage domain decomposition for estimating traffic states from sparse fixed sensors, aiming at scalable inference that stays reliable where data are limited.

I study how transportation networks absorb and recover from disruption. My work quantifies how climate-driven and cyber disruptions propagate through networks, how reliability degrades under such events, and how operations can be designed for faster and more dependable recovery, with a focus on reliability- and recovery-oriented modeling and decision support at the city scale.

I model and mitigate cyber threats to connected and autonomous mobility. This spans route guidance attacks that manipulate driver routing and their behavioral and network-level effects, detection of V2X attacks such as fake emergency messages, and a broader review of cybersecurity needs for next-generation road transportation, with defenses judged by their network-level outcomes.

I combine optimization, game theory, and behavioral modeling to study how travelers route and respond in mobility systems. This includes mean field routing games for connected and autonomous vehicles, analytical network-flow frameworks for traffic under route guidance attacks, and empirical studies of driver behavior such as gap acceptance and surrogate safety measures, connecting individual behavior to network-level performance.