HARI SREENIVASAN: But first: Can robotics and artificial intelligence help improve that rush hour commute you’re facing?
Experts at Carnegie Mellon University think they can by monitoring traffic flow in real time.
Jeffrey Brown has the story from Pittsburgh, part of our weekly series on the Leading Edge of science and technology.
JEFFREY BROWN: You know the frustration. You’re late for work or to pick up your child. You’re driving through city streets, and every block or two, it seems, there’s another red light.
It’s a problem that plagues commuters across the country. In fact, according to a Texas A&M study, the average commuter in the U.S. spends upwards of 42 hours a year at a complete standstill, stuck in traffic.
Eight years ago, traffic problems in Pittsburgh got the attention of a local philanthropist, who gave seed money to Carnegie Mellon University. The idea? To have its robotics experts use artificial intelligence to create a smarter transportation grid that will eventually remake the commute for drivers, cyclists and bus riders.
Almost half of all Pittsburgh commuters drive alone in their cars, so the first priority was road congestion.
Courtney Ehrlichman helps run the program called Traffic21.
COURTNEY EHRLICHMAN, Deputy Executive Director, Traffic 21: The problem in Pittsburgh is like the problems around the country, is that we have this existing infrastructure that was designed many, many years ago, and you really can’t expand it. You really have to optimize the system.
And so we have an opportunity here using these technologies to make our system more efficient and to optimize it, rather than trying to figure out how to build more.
STEVE SMITH, Carnegie Mellon University: We’re going with the flow right now, so we’re getting the greens.
JEFFREY BROWN: Professor Steve Smith thinks he’s discovered a key to optimizing traffic flow. He’s created software that’s currently deployed at 100 intersections in a pilot area of East Pittsburgh.
STEVE SMITH: Conventional signals are pre-programmed. So, it’s really designed for average traffic flows, as opposed to actual traffic flows.
JEFFREY BROWN: So then the trick is to make it as real-time…
STEVE SMITH: Right. We’re watching the real traffic.
JEFFREY BROWN: Smith’s technology uses existing cameras and radar to track how many cars are approaching. Then an algorithm determines how to efficiently move the cars through.
That program controls the lights and also sends information to neighboring intersections.
STEVE SMITH: Essentially, we’re trying to build a signal timing plan that moves all the vehicles from these four different views through the intersection.
JEFFREY BROWN: Without a lot of waiting time.
STEVE SMITH: Yes.
But it also then communicates to its downstream neighbors what traffic it expects to be sending their way.
JEFFREY BROWN: Yes.
STEVE SMITH: So, now downstream neighbor — let’s say it’s this guy, which is a downstream neighbor — is building its own local plan, but now, in addition…
JEFFREY BROWN: Yes. He knows what is coming.
STEVE SMITH: Now, in addition to what it sees in front of it, it has an idea of what is coming down the pike behind what it can see.
JEFFREY BROWN: All right, so here’s one of your boxes.
STEVE SMITH: Yes.
JEFFREY BROWN: The traffic plans are recalculated every few seconds, using a small computer installed at each intersection. Smith says the program has made a noticeable difference.
STEVE SMITH: We pretty consistently get, on average, through the day a 25 percent reduction in travel times, not so much because vehicles are moving faster, but they are stopping 30 percent fewer times.
JEFFREY BROWN: Yes. They’re just continuing to move.
STEVE SMITH: And when they do stop, they’re only idling for like 40 percent less.
JEFFREY BROWN: This summer, the Pittsburgh traffic control plan won an international innovation award.
And even better for the city, the U.S. Department of Transportation has announced it’s giving $11 million to expand the program to many more intersections.
ALEX PAZUCHANICS, Policy Advisor, City of Pittsburgh: Traffic is probably the number one issue that we get complaints about.
JEFFREY BROWN: And that’s because, I mean, that’s life.
Alex Pazuchanics, who works in the mayor’s office, says it may be impossible to make everyone happy, but city officials have been pleased with the program so far.
ALEX PAZUCHANICS: For us, it’s really important that we’re seeing improvements in those corridors, and would like to find ways to expand the use of technology to solve more city issues.
JEFFREY BROWN: In fact, Steve Smith and his team are now working on plans to use smartphones so that cyclists and pedestrians can digitally talk to the intersections to help their commutes too.
Another problem for drivers and the city, potholes and decaying streets. And there’s a project under way here to deal with that as well.
Pittsburgh has more than 800 miles of roads and, like most cities, a limited budget to maintain them. Professor Christoph Mertz has devised software that could help the city better determine where to send repair crews.
So, the idea is pretty simple, right, use the camera that we all carry around?
CHRISTOPH MERTZ, Carnegie Mellon University: Right. I mean, it’s just a smartphone.
JEFFREY BROWN: The phone, which can be mounted on the windshield of any vehicle, records video of the road. Then comes the technological wizardry. Mertz developed software that quickly analyzes the footage, and is capable of distinguishing between small cracks and bigger problems.
So the machine has to think. What is that?
CHRISTOPH MERTZ: Artificial intelligence, it’s like a big word.
And in the past, people thought that, when a machine can play chess, then it’s intelligent. Now, it plays chess, but that’s all it can do. Right? That program doesn’t help us fix the roads.
And so our program will recognize roads.
JEFFREY BROWN: Can’t play chess.
CHRISTOPH MERTZ: But can’t play chess, right.
JEFFREY BROWN: The program also provides simple-to-use color-coded mapping.
So, if I’m a city civil engineer or city manager, how do I use this?
CHRISTOPH MERTZ: You could say, here’s the red, so I’m going to do all the red stuff, OK, but here, this is the main road. Even if it’s just yellow, I want to — I need to fix that, because there’s much more traffic and it’s more important.
JEFFREY BROWN: Mertz and Carnegie Mellon have just launched a private spinoff company called RoadBotics to sell the technology to local governments.
The nearby town of North Huntingdon was first to sign on. One idea is to mount cameras on garbage trucks, which travel every city street once a week.
CHRISTOPH MERTZ: The idea here is, if it’s so inexpensive, you can do it all the time. And so you can address the problems right away.
JEFFREY BROWN: Maybe before it’s a big pothole.
CHRISTOPH MERTZ: Before it’s a big pothole.
So, it’s like for every dollar spent in preventative maintenance, you save $10 in reconstructive maintenance. So, that’s a huge savings.
JEFFREY BROWN: Although Pittsburgh was involved in a pilot project using this technology, it hasn’t yet signed onto a long-term contract.
Alex Pazuchanics explains.
ALEX PAZUCHANICS: Technology is moving very quickly. There’s a risk to making a large investment without necessarily knowing what direction the technology is moving. So, for us, it’s making sure that we’re fast and nimble enough to make decisions as the technology evolves.
JEFFREY BROWN: You mean the technology might be changing so quickly that you’re on the wrong technology?
ALEX PAZUCHANICS: You could very easily be obsolete after just a few months of having the technology out there. We want to be leading edge, not bleeding edge.
JEFFREY BROWN: The hope is, the new technology will also address public transportation, moving buses along as quickly as possible, and giving them priority at crowded intersections.
That could lure more commuters out of their cars, further reducing congestion.
For the PBS NewsHour, I’m Jeffrey Brown in Pittsburgh.