Researchers exploit AI decision making for next generation traffic control

Transport researchers from the University of Southampton are investigating AI solutions for traffic control systems that could transform the performance of road networks in the world’s busiest cities.

The ground-breaking research in partnership with Siemens Mobility is aiming to make next generation traffic systems more efficient, safer and greener for all road users.

Engineers from Southampton’s historic Transportation Research Group (TRG) are tapping into the latest machine learning techniques to create an intelligent model that is adaptable for major cities’ evolving and growing traffic needs.

“Most of the existing algorithms used worldwide to control traffic lights were initially developed in the 1970s when limited real time data was available and the focus was simply on maximising the number of vehicles that could pass through a junction,” Dr Ben Waterson explains. “The world has changed since then with a much greater focus on making things better for all users of the transport system, including cyclists, pedestrians and public transport.

“While the older algorithms have been improved over the years, increasing amounts of real-time data on objects’ movement suggest that newer approaches will be able to outperform them in the future.”

Researchers are focussing on pattern recognition techniques that understand the movement of different sizes of vehicles, cyclists and pedestrians, while also matching real-time sensor data flows to underlying theories of movement. This will then drive the development of decision making AI within the traffic control software which can learn from past decisions to improve future performance.

TRG and Siemens have a long history of working together on urban transport systems, dating back to the creation of the first urban traffic control centres around 30 years ago. “This collaborative relationship with Siemens has achieved much and we will continue to merge cutting-edge theory with practical considerations to achieve real impact on people’s everyday life,” Ben adds.

This latest research builds upon initial findings that were funded through the University’s Web Science Institute Stimulus Fund.

TRG is one of the UK’s longest established and leading centres for engineering-related transport teaching and research. It was founded in 1967 and is based at the University's Boldrewood Innovation Campus.