How AI optimises traffic and reduces the risk of congestion

According to the EU, congested roads cause economic damage of 100 billion euros per year in the member states. But as two research projects now show, the use of artificial intelligence (AI) could make road traffic significantly less congested - for both cars and pedestrians.

How could the current traffic light system with its rigid set of rules and induction loops embedded in the asphalt be further developed? What positive effects for

• the waiting times at the intersection,
• travel times and
• the noise and CO2 pollution caused by traffic jams?

would the use of AI, or more precisely deep reinforcement learning algorithms, bring? And how much could pedestrians - especially the elderly and people with disabilities - benefit from correspondingly shorter waiting times and optimised crossing times?

The trade magazine "industr.com" took a detailed look at two innovative research projects:

• "KI4LSA" by the Institute for Industrial Automation INA of the Fraunhofer Institute of Optronics, System Technologies and Image Exploitation (IOSB), as well as
• KI4PED" by Fraunhofer IOSB-INA together with the company Stührenberg and the partners Straßen.NRW, Stadt Lemgo and Stadt Bielefeld.

You can find the complete article here (in German only):

https://www.industr.com/de/weniger-verkehrsstau-durch-kuenstliche-intelligenz-2638095