Volkswagen software for smart cities Artificial intelligence improves flow of traffic

A pilot project for intelligent urban planning: in Carmel, Indiana, Volkswagen is analysing mobility data to understand the flow of traffic. The automotive manufacturer is using its own Computer Vision software from the production department for this project.

Out of the box: from the factory to the road
The lab is part of Volkswagen’s Software Innovation Center, the Group’s hub of expertise for new IT solutions and innovations. “As inventors, we are always curious and asking ourselves whether it’s possible to go even further,” explains Zarbock. And VW2 went further. In a research project with the city of Carmel in the USA, Volkswagen and Scania researchers are now using the software to analyse traffic. Simply put, the software uses cameras to record the flow of traffic in the city, thus creating an extensive database. This is analysed by artificial intelligence (AI) and helps to optimise the transport infrastructure.
As inventors, we are always curious and asking ourselves whether it’s possible to go even further.

Sunshine, rain, snow: Carmel is the ideal test site
“Software development is an iterative process, which means that we are continuously improving our algorithms and want to push them to their limits,” says Zarbock. Carmel has the perfect conditions for this: “It has all four seasons, meaning we need to deal with bright sunshine, frost, fog, and snow on the cameras.”

The City of Carmel has a track record of exceptional innovativion. It is where the first electric traffic lights in the USA were errected in 1924, and the city now has more than 140 roundabouts – more than any other city in the United States. Carmel has a population just over 100,000 – which is four times larger than it was 30 years ago. This rapid growth in population results in a need for novel urban planning.
Data enables intelligent urban planning
In addition to urban planners, the police force and fire brigade can also benefit from the data. Thanks to updates every ten seconds, traffic is depicted in almost real time. “This is a huge help in specific traffic situations, such as large-scale events or public emergencies,” says Zarbock. This allows first responders to react more quickly to incidents as they happen, such as an unusual cluster of vehicles, sudden gatherings of people, and last but not least, traffic accidents.
The developers are already thinking another step ahead: “We want to equip more streets with technology as a test environment for optimum use of cameras,” says Zarbock. This will make data available for a city-wide traffic model. Examples of what will be measured include routes that cyclists and electric cars prefer to take, to expand cycle paths and the charging infrastructure in a way that meets demand.
