AI research enhancing SIDRA network traffic analysis

A key element of the widely-used analytical road traffic modelling software SIDRA INTERSECTION developed by Akcelik & Associates (A&A) trading as SIDRA SOLUTIONS is the Network model which relies on an iterative technique due to the mathematical complexity of modelling real-life road network operations, especially under high-demand congested traffic conditions.

This project, a collaboration between the University of Queensland research team and A&A, aims to enhance the iterative method in the SIDRA Network Model.

.

Participants

• SIDRA SOLUTIONS
• University of Queensland

Project background

The SIDRA Network Model is an advanced tool widely utilised for traffic analysis and intersection design in networks up to 50 sites. It plays a crucial role in evaluating individual intersection and road network performance, optimising signal timings, and improving overall traffic flow. The SIDRA Network Model uses an iterative approximation method to find an equilibrium solution in modelling traffic interactions between upstream and downstream intersections.

The iterative approach allows for capacity reductions in upstream lanes resulting from blockage by downstream queues, aiming to find a consistent solution between the backward spread of congestion and capacity constraint in upstream lanes. A primary issue with the iterative method is its difficulty in achieving stable and convergent solutions in high demand conditions and in areas of closely spaced intersection where the downstream queues block upstream intersection lanes.

The challenge lies in the interdependent nature of traffic networks (a change in one part of the network causes changes in another). Two fundamental elements of the SIDRA Network Model are:

1. the backward spread of congestion; and
2. capacity constraint to oversaturated upstream lanes.

The two elements are highly interactive as backward spread of congestion blocks upstream lanes and reduces upstream capacity (Figure 1). The reduced capacity in upstream lanes, in turn, reduces downstream arrival flows, queue lengths and therefore decreases probability of upstream lane blockages.

The increasing complexity of urban traffic networks, combined with the growing demand for more accurate and efficient traffic management solutions, highlights the need for improvements in the SIDRA Network Model’s iterative method. Addressing these challenges is essential for providing reliable and timely traffic analyses, optimising signal timings, and ultimately enhancing the overall efficiency of traffic networks, including significant environmental gains.

The two-loop convergence method based on the findings of this research has been incorporated into the latest SIDRA INTERSECTION Version 11.0.3 released recently (Figure 2).

Author contributions

• Mahmud Keblawi led the execution of the project, with responsibilities for all aspects of the research and reporting. This included the development of the conceptual framework and methodologies, the design and implementation of the proposed methods, the coding and computational experiments, the production of figures and visualisations, the analysis of results, and the preparation and editing of the final report.
• Mehmet Yildirimoglu supervised the project, contributed to the development of the framework, conceptual solutions, and methods for enhancing the iterations in the SIDRA Network Model and contributed to the preparation of this final report.
• Rahmi Akçelik prepared the background information on methods used in SIDRA INTERSECTION, contributed to the project definition, verified the data-driven and numerical methods developed during research progress, and contributed to the preparation of the final report. The original iterative method for the SIDRA Network model was developed by Dr R. Akçelik.
• SIDRA SOLUTIONS team members Mark Besley, Harry Cai and Dr Ian Espada contributed to the project definition, review of technical aspects of the research during progress, and preparation of the final report.