When it's that hour before a work day, you know hell's on the loose. But a robotics professor and his AI system will try to ease traffic jams - at least in Pittsburgh.
Carnegie Mellon University professor Stephen Smith said traffic congestion in the U.S. alone is already costing the economy $121-billion a year. Most of this is due to lost productivity.
To make things worse, it also produces around 25-billion kilograms of CO2 emissions.
This is because drivers tend to spend 40-percent of their time being stuck in traffic.
Why? Signals are dumb.
But Smith's startup Surtrac wants to commericalize AI-based
Surtrac's AI system will be changing on the fly. This depends on how severe the traffic jams are and how fast cars need to adapt to them
Pittsburgh's traffic jams: Are they any good?
Results were impressive!
Pilot tests in Pittsburgh revealed that travel time has reduced by 25-percent, and idle times by over 40. The minutes this can spare could be spent doing more work and spending more time with the family.
It also cut emissions by 21-percent! This project may even eliminate street parking and reduce the costs of road widening.
But what is the root of all this? Conventional traffic lights are timed and pre-programmed. Sadly, they only get to be updated every few years so their information get outdated.
Surtrac's AI system coordinate with each other. Sensors and cameras detect light traffic to heavy traffic jams. Their systems can then use this information to plan ahead of time as to what to do.
Unlike other systems, this one is decentralized. This means each signal will make its own decision depending on what it's seeing.
The next step? Why, talking cars of course! Somehow.
Surtrac wants to help give drivers a heads-up about the traffic conditions on-the-go. However, this may be used in public transportation first.