Transportation System Modeling
Analytical and data-driven models of how traffic, travelers, and networks behave.
Why it matters
Automation, electrification, and AI change how vehicles behave, but agencies still decide with system-level questions: how much throughput a bottleneck can carry, what a lane conversion will do to delay, and how travelers will respond. Those questions need models that are both faithful to traffic flow theory and fast enough to use.
What we do
We build analytical models where structure matters, such as queues, discharge rates, and merging, and calibrate them with data. Where data are rich, we use machine learning, and we keep it consistent with traffic flow physics. Recent work ranges from a fluid queue model for heavy congestion to a review of volume-delay functions that pairs classical forms with AI-based ones and comes with open code.
We also make mobility data more useful and safer to share: connected vehicle data for spillover detection and road friction forecasting, and differential privacy methods for publishing transit data.
What we work on
Queueing, bottlenecks, and capacity
Analytical models of throughput at bottlenecks: a fundamental diagram-consistent fluid queue model, multi-state discharge rates for bus-only lane conversion, merging-state roundabout capacity, and a review of volume-delay functions from classical forms to AI.
- A Fundamental Diagram-Consistent Fluid Queue Model for Dynamic Throughput under Heavy Traffic Congestion (TR Part C, 2026)
- Enhancing Traffic State Estimation at Bottlenecks through Improved Demand Modeling: A Greenshields-Grounded Approach (TRR, 2025)
- Analytical Characterization of Multi-State Effective Discharge Rates for Bus-Only Lane Conversion Scheduling Problem (TR Part B, 2021)
Network modeling and demand calibration
Calibrating dynamic origin–destination demand in congested networks, morning commute equilibrium, partitioning networks for macroscopic fundamental diagram-based management, and path algorithms.
- A Sequential Decomposition Framework and Method for Calibrating Dynamic Origin-Destination Demand in a Congested Network (IEEE T-ITS, 2017)
- Morning Commute in a Y-Shaped Traffic Network with Early and Late Arrivals (TR Part E, 2026)
- A Network Partitioning Algorithmic Approach for Macroscopic Fundamental Diagram-Based Hierarchical Traffic Network Management (IEEE T-ITS, 2018)
Connected vehicle data and road conditions
Turning partial connected vehicle data into network intelligence, from estimating turning-movement spillover in near real time to interpolating and forecasting road friction for PennDOT.
- Quasi-Real Time Estimation of Turning Movement Spillover Events Based on Partial Connected Vehicle Data (TR Part C, 2020)
Privacy-preserving mobility data
Differential privacy methods for publishing transit smart card and trajectory data, including how to allocate the privacy budget so the data stay useful for smart mobility applications.
- Optimization of Privacy Budget Allocation in Differential Privacy-Based Public Transit Trajectory Data Publishing for Smart Mobility Applications (IEEE T-ITS, 2023)
- A Differential Privacy-Based Privacy-Preserving Data Publishing Algorithm for Transit Smart Card Data (TR Part C, 2020)
Travel behavior and demand management
How information and incentives change travel choices, built on Dr. Hu's work at Metropia on incentive-based active demand management and on telematics-based driving risk for usage-based insurance.
- Will Information and Incentive Affect Traveler's Day-to-Day Departure Time Decisions? An Empirical Study of Decision Making Evolution Process (IJST, 2020)
- Triggering Behavior Changes with Information and Incentives: An Active Traffic and Demand Management-Oriented Review (ATPP, 2019)
- Behavioral Responses to Pre-Planned Road Capacity Reduction Based on Smartphone GPS Trajectory Data: A Functional Data Analysis Approach (JITS, 2019)
Representative work
A Fundamental Diagram-Consistent Fluid Queue Model for Dynamic Throughput under Heavy Traffic Congestion
A fluid queue model of dynamic throughput under heavy congestion that remains consistent with the fundamental diagram of traffic flow.
Analytical Characterization of Multi-State Effective Discharge Rates for Bus-Only Lane Conversion Scheduling Problem
Characterizes multi-state effective discharge rates and uses them to schedule when a lane is converted to bus-only operation.
Analytical Approximation and Calibration of Roundabout Capacity: A Merging State Transition-Based Modeling Approach
An analytical approximation of roundabout capacity built on merging state transitions, with a calibration method and public code.
A Review of Volume-Delay Functions Integrating Traditional Models and Emerging AI Technologies
Reviews volume-delay functions from classical forms to AI-based models, with open code for comparing them.
Data and resources
Sponsored projects
All projectsFriction Factor Interpolation and Forecasting
Pennsylvania Department of Transportation
2026
PIAI-Based Prediction Models for Transportation Infrastructure Asset Management Data Hub – Phase I
U.S. Department of Transportation (NCTriDurLE UTC)
2022–2023
PSU PIPreparing Multidisciplinary Datahub for Digital and Intelligent Infrastructure System – Preliminary Phase
Center for Intelligent Infrastructure, Missouri S&T
2020–2021
Co-PIConnectSafe: Connect Usage Based Insurance, Transportation Big Data and Socioeconomic Behavioral Research in Building a Safe-Driving Community
Smart Living Research Center, Missouri S&T
2018–2019
PIUtilizing DynuStudio-DynusT Dynamic Traffic Assignment Software Suite to Model the Fuzhou Transportation Network
Fujian University of Technology
2016
PIExploring Emerging Traffic Data Collection for Transportation Planning and Operations Using the Metropia System
Pima Association of Governments
2015–2016
Co-PIPay-As-You-Drive-And-You-Save (PAYDAYS) Insurance Actuarial Study
Federal Highway Administration
2013–2016
Co-PI and Project Manager
Recent publications
All 22 in this thrustMorning Commute in a Y-Shaped Traffic Network with Early and Late Arrivals
A Fundamental Diagram-Consistent Fluid Queue Model for Dynamic Throughput under Heavy Traffic Congestion
Enhancing Traffic State Estimation at Bottlenecks through Improved Demand Modeling: A Greenshields-Grounded Approach
A Review of Volume-Delay Functions Integrating Traditional Models and Emerging AI Technologies
Analytical Formulation of Autonomous Vehicle Freeway Merging Control with State-Dependent Discharge Rates
A Multi-State Merging Based Analytical Model for an Operation Design Domain of Autonomous Vehicles in Work Zones on Two-lane Highways
Optimization of Privacy Budget Allocation in Differential Privacy-Based Public Transit Trajectory Data Publishing for Smart Mobility Applications
Analytical Approximation and Calibration of Roundabout Capacity: A Merging State Transition-Based Modeling Approach
Analytical Characterization of Multi-State Effective Discharge Rates for Bus-Only Lane Conversion Scheduling Problem
Will Information and Incentive Affect Traveler's Day-to-Day Departure Time Decisions? An Empirical Study of Decision Making Evolution Process
Work with us
Our models help agencies evaluate operational strategies, such as lane conversions, before they build them, and get more out of the connected vehicle data they already receive.
Email Dr. Hu