The lights are on

At a busy intersection in urban Hamburg, Germany, researchers have been trialling a new kind of traffic light. That might not sound very exciting, but their unique method of cutting stop-times at red lights and keeping the traffic moving is based on an entirely new way of thinking about computer systems.

Normally, traffic light systems use either fixed timer controls or a centralised control system. Both are based on a technique that tries to keep lights green at certain times, so that traffic flows smoothly and pedestrians can cross roads safely. Such systems can be highly sophisticated, but they aren’t always effective – as anyone who has tried to cross a major urban area at rush hour can testify. Fixed timers are flawed because they do not respond to the volume of traffic. Centralised systems are more responsive, but not optimal. Drivers still end up stuck at red lights when they could be moving forward, wasting time and fuel and creating unnecessary pollution.

The experimental Hamburg traffic lights have done a much better job, using a technique inspired by what is known as “organic computing”. An organic approach to computing systems is based on the idea that we will soon be surrounded by large collections of autonomous systems, of which traffic lights are a good example. These systems, equipped with sensors and actuators, are aware of their environment, can communicate freely, and organise themselves in order to perform the actions and services that seem to be required. Organic traffic light systems can collect their own information about traffic flows and work out for themselves whether they should change the lights to red or green.

According to the team behind the Hamburg traffic lights – Holger Prothmann of the Karlsruhe Institute of Technology and colleagues there and at Leibniz Universität Hannover – organic computing can model even very complex systems. In recent years, it has emerged as a possible solution to a wide range of problems involving complex, autonomous systems that have sensors and controllers, they argue in a recent edition of the International Journal of Autonomous and Adaptive Communications Systems.

“The environmental and economic importance of traffic control systems combined with the distributed nature of traffic nodes and their constantly changing traffic demands make traffic light control an ideal test case for organic computing approaches,” explains Prothmann.

The Hamburg test uses industry-standard traffic light controllers, adapted to have an observer/controller computer architecture that allows the traffic light to respond to traffic flow and to pass on information to the other traffic lights on neighbouring roads. The results show that the average number of vehicle stops can be cut significantly, delays avoided, and journey times reduced, all of which has benefits for drivers, pedestrians and city dwellers.