Fifty years ago the Federal Aid Highway Act of 1956 created the Interstate Highway System, the extensive network of roadways that helped bring prosperity and development to many areas of the country. This weekend, millions of drivers will celebrate the anniversary by sitting in traffic on many of those same roadways. Traffic is enormously costly: in U.S. urban areas alone, it causes almost 4 billion hours of delay and wastes more than 2 billion gallons of fuel per year, according to the Texas Transportation Institute in College Station, TX.
Can’t technology help? Technology Review asked Kara Kockelman, associate professor of transportation engineering at the University of Texas, Austin, why 50 years after “Ike’s Autobahns” were built, we’re still stuck in traffic.
Technology Review: We’ve seen traffic detectors like cameras and roadway loops for years. How well are these kinds of basic traffic technologies working?
Kara Kockelman: Freeway cameras are largely used for surveillance of traffic back-ups and to guide emergency response. Detectors are primarily for counting vehicles, including limited metering of on-ramp traffic. Some of the resulting travel-time information is being provided to travelers in places like Seattle, Minneapolis, and Los Angeles. But while information is helpful to those seeking to avoid some congestion, these basic technologies are not moving traffic much, if any, faster because travelers receiving the information already are on the road. Right now, radio-broadcast traffic reports may be just as effective for relaying news of such events to drivers.
TR: What about those separate high-occupancy or HOV lanes, for drivers who have two or three or more people in the car? Are they working?
KK: Pravin Varaiya’s team’s study [at the University of California, Berkeley] of detector data along California highways indicated that HOV lanes actually worsen traffic flow by reducing lane-choice flexibility and allowing slower HOV vehicles to determine the HOV lane’s speed. And cars seeking to enter left-side HOV lanes create traffic-weaving, another source of delays. To me, this suggests that pricing is needed, so that people can decide how much they are willing to pay to get somewhere faster.
TR: Would it help to have cars “talking” to one another, relaying traffic information, and making more efficient use of the roadway space?
KK: In terms of the highway system, most people don’t see inter-vehicle communication for more efficient road space use taking off anytime soon. Most people feel that the liability issues are too serious. However, some of the associated technology is being put to use in other, often safety-related applications, such as adaptive cruise control, in which a car with cruise control senses a car ahead and can slow down. And automated vehicle guidance systems may result in caravans of driverless trucks, all following a single driver.
TR: Presumably, Global Positioning System technology could be a key enabler of the entire traffic system?
KK: GPS on trucks for tolled travel is moving forward in Germany, and Oregon’s Department of Transportation is experimenting with GPS-based mileage pricing, offering gas-tax breaks according to travel distances, as recorded by position data. They aim to soon introduce a variable-pricing form of this, where peak-period driving will cost 10 cents per mile within the Portland area, again, recorded by GPS position data. I’d like to play a role in that work, by estimating the benefits (and costs) experienced by everyone in the region. Many states hope that vehicle manufacturers will pursue inclusion of GPS in all vehicles, to tame that technology’s presently high fixed costs while permitting relatively seamless tolling.
TR: You propose to evaluate the impacts of what you call “credit-based congestion pricing.” What’s that?
KK: Credit-based congestion pricing is a policy wherein tolls rise with traffic demand, thus keeping traffic moving, and distribution of travel budgets ensures a reasonable level of access for everyone. For example, the first 100 miles a vehicle travels during peak periods each month would be “free.” But after that, the driver is on his or her own, paying tolls out of pocket, via a transponder account. And the tolls would be higher during times of congestion. This incentive structure optimizes demand by allowing tolls to vary with congestion, on each link in the system across all times of day.
TR: Is there any evidence that this could work better than previous technology efforts?
KK: Our models of the Dallas and Austin systems indicate that such policies can eliminate recurrent forms of congestion while enhancing the lives of most residents. People are very interested in this form of pricing policy, as it makes good sense and resonates with the public, thanks, in part, to its equity implications – no one is priced off the network. San Jose, California’s transportation authority incorporated my ideas into a federal highway pilot program proposal, and I hope that work will be going forward.
TR: Meanwhile, is anything significantly increasing capacity on highways?
KK: Yes. We’re seeing more aggressive driving patterns, and much higher speeds. This increases capacity “naturally” from 2,000 vehicles per hour per lane to more than 2,400. Of course, at higher speeds, injury severity and survivability of crashes becomes a serious issue. Vehicle improvements are occurring, certainly; but, ultimately, there is only so much auto manufacturers can do to save us from the speed-squared law of kinetic energy.
How AI is reinventing what computers are
Three key ways artificial intelligence is changing what it means to compute.
These weird virtual creatures evolve their bodies to solve problems
They show how intelligence and body plans are closely linked—and could unlock AI for robots.
A horrifying new AI app swaps women into porn videos with a click
Deepfake researchers have long feared the day this would arrive.
Surgeons have successfully tested a pig’s kidney in a human patient
The test, in a brain-dead patient, was very short but represents a milestone in the long quest to use animal organs in human transplants.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.