low emission zone
The city of Antwerp is getting a boost from foreign drivers with an electric car. Many of them receive a fine if they enter the low-emission zone (LEZ), even if their vehicle does not emit any harmful substances. The city is looking to AI to solve that problem, recognizing vehicles instead of license plates.
Since February 2017, Antwerp has been banning the most polluting cars from the city center and on the Left Bank. Camera systems check incoming cars by looking up their license plates in a database that knows the emissions per car.
But that database only contains Belgian and Dutch license plates. Drivers from other countries must register their car in advance and specify the emission standard. Anyone who does not do this and still enters the LEZ risks a fine of 150 to 350 euros.
The city council is struggling with these automatic fines, because sometimes it concerns an electrically powered vehicle, or it is a car with a fuel engine that is clean enough. “It causes a lot of bad blood among those drivers abroad, it is experienced as a tax on bullying,” says Antwerp Alderman for the Environment Tatjana Scheck (Vooruit). According to her, the vast majority of fines currently concern vehicles with a foreign license plate.
European system
That is why the city is working on a solution. For example, she launched a market survey to do more with artificial intelligence than just reading license plates. “The city wants to explore techniques to detect and recognize the vehicles themselves in terms of type and age. We could still recognize compliant cars via vehicle recognition and therefore automatically grant access, without a fine being issued. That could take away a lot of frustration.”
The system should be able to recognize not only electric vehicles, but also more recent fuel vehicles that meet the applicable emissions standard (which is periodically tightened).
The solution with AI cameras is not preferred. “Just as we have Dutch license plates in our database, we would have preferred a European system of data exchange of license plates, which we have always strongly insisted on. That would be much more efficient.”
According to the alderman, on the other hand, analyzes can also be done with AI vehicle recognition: “Think of mapping traffic flows in the city, how many vans are driving around, the share of heavy traffic, etc.,” Scheck concludes.
There is currently no clarity about when a more fine-grained detection system can be introduced.
