A data-driven approach for network-wide modelling and control of urban network

This research aims to model a large-scale urban network, consisted of thousands of short links and traffic signals, using a limited set of data collected from a combination of fixed and mobile traffic sensors. The required data for this study will be provided by the AMA Eco-drive trajectory data and Edmonton citywide Dynameq model. A general framework for describing a large traffic network will be developed using machine learning/artificial intelligence tools. Based on the developed traffic model, a congestion pricing scheme to optimize tolling prices and zones will be developed. Finally, a scenario analysis of the different tolling and congestion levels will be conducted. This study enhances computational efficiency and realism aspects of traffic modelling by describing large traffic networks using a limited set of links and vehicle trajectories.

Faculty Supervisor:

Lina Kattan

Student:

Partner:

Alberta Motor Association

Discipline:

Engineering

Sector:

Administrative and support, waste management and remediation services; Finance and Insurance; Other services (except public administration)

University:

University of Calgary

Program: