The University of Auckland

Project #32: Applications of machine learning in transportation technology - Part 2: Traffic Control Systems.

Back

Description:

 

Auckland is a large, continually growing city, and as a result those who commute suffer the financial, social and environmental consequences of excessive traffic every day. Therefore, the motivation of this project is to improve the existing infrastructure to reduce congestion, travel time, and as a result carbon emissions. The two stages of this project are: 1. The simulation of traffic within a specified area within Auckland, and 2. Algorithm design of traffic lights to improve traffic flow. We are responsible for the research and development of the Traffic Control System algorithms.

The New Zealand Transport Agency (NZTA) has currently implemented the Sydney Coordinated Traffic Control System (SCATS) in Auckland, which was developed in the 1970s. This system lacks the flexibility to self-adapt to the rapidly changing transportation landscape, and has since been surpassed by newer, state-of-the-art traffic control systems.

 We aim to take a data-driven approach to improve on SCATS by using machine learning to develop a more sophisticated traffic control system. Research will be undertaken to determine modern algorithm designs that are suitable for this application, from which we will develop one or more machine learning architectures. The simulation will allow us to analyse the performance of these algorithms in comparison to the current system.  

Outcome:

 

 

 

Research Components:

●      Research of machine learning algorithms in transportation applications

●      Research of modern deterministic traffic control systems

●      Researching suitable traffic simulation methods and tools

●      Development of a traffic control system that utilizes machine learning.

 

Prerequisites

COMPSYS 302

Specialisations

Categories

Supervisor

Co-supervisor

Team

Lab

Lab allocations have not been finalised