Dr Wei Huang

Postdoctoral Research Fellow
Department of Civil & Environmental Engineering

The Hong Kong University of Science and Technology

Dr Wei Huang is currently a Postdoctoral Research Fellow of the Department of Civil and Environmental Engineering at The Hong Kong University of Science and Technology (HKUST). She received her PhD degree in traffic engineering from KU Leuven, Belgium, in 2015. She obtained her Bachelor's degree and Master's degree in Traffic Information Engineering and Control from Tongji University. She was awarded Shanghai Outstanding University Graduate in 2007. Wei Huang has been involved in several research projects, including Program of the National Natural Science Foundation of China, National High-Tech R & D Program, Flemish Government Program, and Strategic Public Policy Research Funding Scheme by the HKSAR Government. Her research interests include Intelligent Transportation Systems, transportation network modeling, control and optimization, reliable transportation network design, and iterative learning control. She has published over 20 journal and conference papers in transportation research, including TRC, TTRB, JAT, ISTTT, and IEEE-ITSC.

Enhancing Anticipatory Network Traffic Control with Iterative Learning Optimization

In urban traffic networks, travel delay at signal-controlled intersections is a main component of the travel time experienced by travelers. Effective optimization of traffic signal control is an important tool to address the growing travel demand and alleviate network traffic congestion. Among others, anticipatory traffic control determines signal timings to optimize network-wide objectives, e.g. minimize total travel times, taking into account travelers' route choice responses and the resultant equilibrium flows. The route choice response is usually predicted through a response function known as traffic assignment model. However, the model-reality mismatch unavoidably brings suboptimal performance, which is characterized by unexpected congestion to the real-life system. Therefore, there is a need for introducing corrective measures to deal with the model approximation errors and hence to achieve an effective control in reality. In this study, a novel method to support control decisions for practical applications is introduced. We take advantage of the fact that routine traffic patterns repeat from epoch to epoch (for instance from day to day or week to week), allowing one to learn from observing the true traffic patterns to compensate for errors in the response model. This study demonstrates an effective optimization of anticipatory traffic signal control, which integrates the model-based control design with data-driven learning techniques. The main objective of this study is to elevate the traffic system to its best achievable performance.