Artificial system

Less waiting and safer: the results of applying artificial intelligence to traffic lights

Traffic management in a city is one of the most complicated tasks in mobility. In this work, the management of traffic lights and their timetables It seems decisive when it comes to meeting all needs. A problem that artificial intelligence can solve.

In a small German town called Lemgo, located near Hanover, a pilot project has been started in which artificial intelligence traffic lights are used, with the aim of improving traffic. The first data confirms that this system is working and that journeys are now faster.

The tests are funded by the German Federal Ministry of Transport and Digital Infrastructure under the name of “KI4LSA” and “KI4PED”attended by industrial automation specialists from the Fraunhofer Institute for Optronics, System Technologies and IOSB Image Exploitation.

An improvement between ten and 15%… for now

These smart traffic lights are nothing more than the conversion of these devices from a busy city intersection. High-resolution cameras and radars have been installed there which study the volume of traffic passing through them, the average speed of vehicles and their waiting times. Thanks to algorithms designed for its own learning, the traffic light improves the regulation of traffic over time.

The first results of the project (which ends this summer) ensure that it has reached improve traffic flow between 10 and 15%, in particular by subtracting the waiting time that vehicles spend standing still. But these results can be improved according to the scientists, who hope to reduce the waiting time by up to 30%.

In addition to shorter journeys in town, the researchers add that these smart traffic lights reduce noise and CO2 emissions and eliminate traffic congestion in town which, according to the European Union, is estimated to cost the States of 100,000 million euros. per year, which the researchers themselves collect in their studies.

Also for pedestrians

So far, the project we talked about operates under the name “KI4LSA”. However, the pedestrians they will also benefit from the possible benefits of this research, through the “KI4PED” project, which focuses on the latter.

The objective is the same as for cars: reduce the waiting time for pedestrians, analyze the time needed to reach the traffic light and keep it open long enough for it to cross safely. This can be particularly useful for the elderly or those with reduced mobility, for whom some of the more common obstacles in their path can be removed.

However, survey officials point out that so far a LiDAR radar to detect pedestrians, which are seen as a cloud of points by the system, in order to guarantee your confidentiality. And although they are not located individually, the point on the map that represents a pedestrian can indicate the number of people with motor difficulties who may need to cross the street, determining their speed.

The implementation of this system can help to significantly improve the mobility data of the city and that the traffic lights themselves adapt to the needs of each moment. Indeed, the use of artificial intelligence should improve traffic fluidity by 30%, but also reduce any incidents at the crossroads by 25%.

Traffic lights, permanently signaled

It is not the first time that traffic lights have been singled out as the main culprits of poor traffic management or one of the big moles of urban mobility. Neither the application of algorithms based on artificial intelligence This is one of the first solutions that were put on the table.

Here, connectivity, the development of new networks and the implementation of 5G have a lot to say. Audi, for example, already has a system in which its cars can exchange information with the next traffic light in the city of Düsserdolf. Your dashboard shows whether it’s open or closed and how much time is left before it turns green in the latter case.

Tests were also carried out in Malaga last year with a smart traffic light and a fully autonomous city bus. In this case, the objective was to study the work to be done by a traffic light in a fully autonomous vehicle environmenthow it could handle each situation and improve the flow of traffic in its environment, taking into account the type of vehicles used or the volume of traffic.