Razy Zubi
Postdoctoral Associate
Affiliation: NYU Abu Dhabi
Education: BSc Technion–Israel Institute of Technology; Direct PhD, Technion–Israel Institute of Technology
Research Websites: Center for Interacting Urban Networks (CITIES)
Research Areas: Traffic flow modeling; Urban traffic signal control; Large-scale network control; Intelligent transportation systems; Traffic estimation and monitoring; Max-Pressure control; Reinforcement learning for transportation; Adaptive traffic control; Automatic control theory
Razy Zubi is a Post-Doctoral Associate at the Center for Interacting Urban Networks (CITIES) in the Division of Engineering at New York University Abu Dhabi. He received his Direct PhD in Civil and Environmental Engineering from the Technion – Israel Institute of Technology in 2026, and previously earned two BSc degrees with Cum Laude honors in Structural Engineering and Transportation Engineering from the Technion.
His research focuses on traffic flow modeling and control, urban traffic signal control, large-scale transportation network management, traffic estimation and monitoring, and intelligent transportation systems. His work combines transportation engineering with automatic control theory, with particular emphasis on adaptive and decentralized traffic signal control, Max-Pressure control.
Zubi is the author of publications in Transportation Research Part C and major transportation and control venues, including IEEE ITSC, IFAC CTS, ECC, CDC, and TRB. His recent work includes studies on travel-time-based and queue-based Max-Pressure traffic signal control, stability of decentralized Max-Pressure control, and data-driven urban traffic management. Before joining NYU Abu Dhabi, he served as a Visiting PhD Student at Chalmers University of Technology and as a Teaching Assistant in Civil and Environmental Engineering at the Technion. He has also received several academic excellence awards in transportation engineering and civil engineering.
Summary of Research
Zubi's research aims to develop a Global Urban Digital Twin for reconstructing and predicting urban traffic states under sparse sensing conditions. The core objective is to estimate high-resolution link-level traffic states, turning ratios, dynamic origin–destination demand, and latent disturbance fields using limited probe trajectories, loop detectors, and semantic urban information.
The project is based on semantic–physics fusion: urban features such as land use, building typology, curbside parking and loading activity, public transport infrastructure, pedestrian activity, enforcement patterns, events, and incidents are treated as physics-relevant priors that affect capacity, friction, route choice, and network performance. Instead of treating irregular traffic patterns as random noise, the framework models them as structured disturbances that can be inferred, interpreted, and predicted.
The main research directions include semantic physics-informed state reconstruction, semantic-gravity turning-ratio estimation, activity-regularized dynamic OD reconstruction, and predictive friction and risk fields. The long-term goal is to enable transferable, low-complexity urban digital twins that can support scalable monitoring, proactive traffic operations, signal control, perimeter control, and planning decisions in cities with limited sensing infrastructure.