Science and Technology

Transport predictive solution Stage 2: AI and real-time simulation

This project aims to offer a real-time decision support tool for traffic operations centres to predict congestion on the network, quickly assess the impact of unplanned events and evaluate the mitigation potential of several possible responses.

Such a solution will help reduce congestion, especially in non-recurrent situations, and significantly increase travel time reliability.

The use of tools to facilitate longer-term prediction of how transportation networks will perform in the future is a well-established practice in strategic planning by transport authorities. Tools to support day-to-day operations, relying on short-term predictions, are in their infancy, especially in Australia.

Particular objectives to enhance short-term prediction performance are:

  1. Smart sensing for enhanced travel demand estimation; and
  2. Artificial intelligence (AI) and machine learning (ML) for calibration against much larger real-time datasets

The WA node will focus on (2) developing and testing improved model calibration capability for both live and offline models, ensuring prediction accuracy for any hour of the day, seven days a week.

This research proposes to improve model calibration and the accuracy of 24 hour/ 7-day models (live and offline) for not just the AM and PM peaks but any hour of any day. The research results will be tested in a WA Aimsun Live network pilot model, developed as part of the more comprehensive project. Further evaluation and performance accessibility of tools developed in this research will be performed in QLD Aimsun Live network model.

Participants

See the full list of iMOVE projects here

Project background

This project agreement forms part of a more comprehensive program of works aiming to enhance real-time traffic prediction. The program of works will be delivered in 2 stages and will include several project agreements (number to be determined):

  • Stage 1 is the Aimsun Live pilot project for WA (Project 1-024) and for QLD (Project 1-026)
  • Stage 2 are the Research and Development projects (for WA and the upcoming Project 1-027 for QLD)

Once the individual components of the research have been developed and tested in the two participating geographic nodes, the study will be integrated into a new version of Aimsun Live which will be applied and evaluated in each of the participating jurisdictions.

WA Node – research need

Main Roads Western Australia is responsible for delivering and managing a safe and efficient road network in WA through operations using technology and other management and operational measures to optimise real-time management and operation of the network.

The intended research around Artificial Intelligence and Machine Learning will improve the calibration of Live Aimsun models. It will also benefit offline models, which are already used within the Operational Modelling and Visualisation (OMV) team in Network Operations.

This research is being undertaken to address two problems related to real-time operations and traffic modelling:

Real-Time Traffic Operations (RTTO): Incident Impact Analysis

When an incident occurs on the road network, Main Roads Western Australia currently relies on the expertise of its employees to identify the likely traffic impacts and how to best manage the situation. Whilst depending on employee experience is beneficial; there are risks and limitations.

Incidents and the traffic and other environmental conditions in which they occur may be different, so the resulting network impacts may not be the same with each incident. The degree of viability, the rapid development of incidents, and the number of possible outcomes can stretch their cognitive abilities, so experience alone may not be sufficient to produce the best outcome to non-recurring situations and incidents.

Model calibration and scenario testing

The OMV team is developing microsimulation models for some of the state routes under Main Roads’ management. These are corridor models (as opposed to network models), and they provide a platform for internal testing of scenarios for the Transport Portfolio and local governments.

These corridor models were developed to be used in projects, appropriately updated rather than built from scratch, to reduce the expenditure of taxpayers’ money.

However, once these corridor models are available, they will comprise AM and PM peak-hour models that will need to be kept updated and require regular traffic surveys, which are both time-consuming and costly.

The OMV team is also considered the custodian for the CBD Aimsun mesoscopic model, owned by the WA Department of Transport (DoT). Negotiations between Main Roads Western Australia and WA DoT are currently taking place to establish the governance and maintenance of the model (annual or biannual basis), including the methodology of how the model should be calibrated.

This project provides several specific benefits for Main Roads Western Australia, including:

  • Real-time predictive modelling capability is the next stage towards Network Operations’ vision of “Predict 20 minutes ahead, act in 5 and change the future”. This project, therefore, is in close alignment with and supporting this vision.
  • Trialling a leading real-time predictive modelling software could take significantly less time to deploy than alternatives and be more efficient and reliable.
  • This is a valuable learning opportunity for the OMV and RTTO teams within Network Operations. It will potentially inform any future procurement process for the acquisition of real-time predictive modelling software.
  • The selection of Perth CBD and the surrounding areas is strategic. It provides an opportunity to model the potential road network impacts of large-scale events undertaken in the area.

Project objectives

The objectives of this research component are to develop Aimsun modules that:

  • Enable the identification of data outliers and repeated data errors to alert human modellers
  • Provide a robust solution for calibration of inputs and parameters for supply, demand, and driver behaviour
  • Facilitate identifying traffic patterns, detecting changes in pattern and associated optimal parameter set of demand and supply for each pattern
  • Estimate confidence levels of traffic predictions

In addition, the WA Node will support the integration and testing of the solution in WA and QLD.

More from iMOVE

Most Popular

ADVERTISEMENT

To Top