Non-recurrent events such as road-accidents can severely affect traffic movement in dense urban areas and lead to increased travel times and intervention costs to alleviate congestion. Often, traffic management centres face the difficulty of clearing an accident, redirecting the traffic in the affected area and/or replacing affected public transport modes with alternative solutions without anticipating the impact of their response actions. This project aims at using traffic simulation modelling reinforced by data-driven solutions in order to learn from previous similar recurrent incident situations, evaluate and predict the impact of new arising accidents and compare the outcome of possible intervention scenarios.
This project will help traffic management centres from large urban areas to evaluate the severity and impact of recent accidents through either a data-driven module if similar traffic patterns and conditions have occurred previously in the network, or through a reinforced traffic simulation if the accident is non-recurrent. The project is jointly applied with counterparties from Australia university and research institute through the ARC funding scheme.
The researchers from both sides will work together towards the following objectives:
1) Incident detection and classification
2) Multi-modal impact analysis
3) Data-driven traffic simulation
4) Response plan generation
5) Real data evaluation
The proposed research scope aligns with the priorities of developing more efficient and reliable transportation systems for both Singapore and Australia’s new changing urban conditions.
Project Title: Data-driven Traffic Analytics and Simulation for Incident Impact Analysis and Management
Host University: Nanyang Technological University and National University of Singapore
Principal Investigator: Dr. Mo Li, Associate Professor, School of Computer Science and Engineering, College of Engineering, Nanyang Technological University
Co-Investigator: Dr. Gary Tan, Associate Professor, Department of Computer Science, School of Computing, National University of Singapore
Collaborators: Prof. Timos Sellis, Professor, Swinburne Data Science Research Institute, Swinburne University of Technology; Dr. Adriana-Simona Mihaita, Research Scientist, Data61, Advanced Data Analytics in Transport, CSIRO (Commonwealth Scientific and Industrial Research Organisation); Dr. Kai Qin, Associate Professor, Faculty of Science, Engineering & Technology, School of Software and Electrical Engineering & Data Science Research, Swinburne University of Technology