This project aims to improve travel demand calibration and accuracy of 24 hour/ 7 days network simulation models (live and offline) for the AM and PM peak hours, in fact any hour of any day. Such a solution will help reduce congestion, especially in non-recurrent situations and significantly increase travel time reliability. The research results will be tested in TMRs Aimsun Live model, which is also being developed as part of the wider project.
Traffic simulation models can now simulate and predict traffic flows of entire metropolitan areas in real-time, empowering network managers and transport authorities to predict and alleviate road congestion under recurrent and non-recurrent events. As the reliability of simulation models depend on the quality of the inputs and the calibration and validation of the model parameters, the specific objectives of the iMOVE project are to enhance short-term prediction performance by utilising:
- Smart sensing for enhanced travel demand estimation and prediction; and
- Artificial intelligence (AI) and machine learning (ML) for calibration against much larger real-time datasets
The Queensland node of the iMOVE project will focus on modelling computationally efficient and reliable travel demand inputs and driving behaviour parameter adjustment for a real-time traffic prediction and management in urban networks using advanced data science techniques on real traffic monitoring data.
The project will integrate and evaluate the solutions for (1) and (2) in Aimsun Live and demonstrate the resulting performance in both collaborating jurisdictions in Australia (Queensland and Western Australia).
This research is being undertaken to address the following problems related to real-time operations and traffic modelling: Effective traffic management solutions need increasingly robust and reliable traffic state estimates and predictions derived from simulation models.
The modelling process involves calibrating and predicting travel demand, using conventional data such as historical travel demand patterns and traffic counts. Traffic monitoring data needs to be intelligently incorporated into the demand estimation and prediction frameworks to enable more reliable travel demand information development and deployment. This will provide the basis for predictive solutions, either real-time traffic prediction, such as Aimsun Live, or traditional offline, mid to long-term transport planning models, such as Aimsun Next.
Today’s road networks are better equipped than ever to provide live monitoring of the traffic via sensors (such as inductive loop detectors, Bluetooth MAC Scanners, etc.). In addition to fixed-location sensors, road agencies are also collaborating with third parties such as taxis, HERE, and Intelematics to further complement their road monitoring data sets.
With the availability of such extensive and heterogeneous traffic datasets, opportunities extend traffic congestion situational awareness from traditional flow and speed characteristics to advanced (partial) vehicular trajectories dynamics. It opens avenues to explore advanced data science algorithms for detailed modelling of traffic congestion dynamics, understanding demand distribution and drivers’ route choice behaviour in the network.
Several benefits to road agencies (TMR) can be achieved by advancing data science techniques on the big traffic datasets for use in real-time simulation, including (but not limited to):
- Trialling one of the industry-leading real-time traffic predictive modelling solutions to confirm the most time-efficient methodology for model network maintenance and operations. This is a valuable learning opportunity for transport modelling, real-time network operations and optimisation teams.
- Real-time predictive modelling capability progresses towards Network Operations’ vision of informed and intelligent control compared to the current reactive control. This project directly aligns with this vision.
- The research application will significantly enhance the operators’ capabilities for real-time and predictive traveller information and management, especially during non-recurrent (incident) and compounded traffic conditions, the time when such information is most critical in averting catastrophic congestion. This extends
- beyond the existing data-driven-only models, which are limited by pattern matching in the historical profiles and unable to determine causality and driver responses.
- The selection of South-East Queensland and the surrounding areas are strategic. It provides an opportunity to model large-scale events’ potential road network impacts (including Brisbane 2032 Olympics and Paralympics games).
This QLD node R&D project aims to develop and integrate advanced data analytics for traffic demand calibration to improve traffic simulation prediction quality and ensure reliable evaluation of traffic management scenarios with Aimsun Live.
The R&D project will exploit the availability of extensive traffic monitoring datasets and utilise advanced data science techniques to address the calibration needs for real-time traffic simulation models. The development and testing of the project solutions will be first evaluated in TMR’s Aimsun Live testbed.
The objectives of this research are to develop Aimsun modules to:
- Establish a library of traffic states and OD matrices derived from a real-time data source
- Improve the short-term traffic demand matching model for real-time simulation in Aimsun Live
- Enrich the route choice calibration for simulation with Aimsun Live for the pilot system
- Deploy the developed solution into the pilot environment
- Demonstrate the transferability of the QLD node solution into the WA node
More from iMOVE