Tomorrow's AI traffic lights will predict when you want to cross the street
Researchers in Austria have developed a smart traffic light that analyzes pedestrians' behavior, spots people who are planning to cross the street, and stops the traffic for them without the need to press a button and wait.
The system, which is due to hit the streets of Vienna next year, uses cameras that monitor an area measuring eight by five meters. It monitors the movements of people walking near the road, and recognizes an intention to cross within one second.
"After two seconds the estimation becomes reliable," says Horst Possegger from Vienna's Institute of Computer Graphics and Vision, which created the system.
- Driverless cars explained: everything you need to know
- Your complete guide to electric vehicles
- When will self-driving cars earn our trust?
If a large group of people are showing an intention to cross, the light can hold the traffic for longer to give them all time to reach the other side safely.
It can also make accommodations for people with wheelchairs, prams or pushchairs, who will need a little extra time.
Stop, scanner time
You might not like the idea of cameras on public streets mapping your every move and predicting your actions, but Possegger and his team are keen to point out that all the monitoring and calculations happen within the light itself, and no data is transmitted to external servers. It doesn't pick up individual people's features, either, instead using geometric data.
The lights also have a system to report faults automatically, so pedestrians aren't left stranded by the roadside if the camera fails. “This is a double safeguard," says Possegger. "The system was developed in such a way that it can work round-the-clock even in a harsh environment and can also deal with voltage fluctuations.”
If it proves successful, we could see the smart lights rolled out in other countries too, where they'd keep the impending tide of autonomous cars in check.
Via Engadget
from TechRadar - All the latest technology news http://bit.ly/2WaDhr4
No comments