The Road Ahead
Plenty of curves lurk on the “intelligent highway.”
A long route 90 near San Antonio, brake lights are starting to flash. Twenty-five kilometers away, Sam Mendoza is sipping coffee in the region’s “TransGuide” traffic management center,where 16 wall-mounted television monitors display scenes from some of the 109 video cameras peering at 100 kilometers of area highways. Suddenly, a beeping sound signals that traffic on Route 90 has slowed below 40 kilometers per hour, as sensed by some of the 1,700-odd magnetic-loop detectors embedded in the region’s roads.Mendoza seizes a computer mouse and zooms a highway-mounted camera toward the problem spot; soon his monitor reveals two bearded men in the breakdown lane struggling to fix a pickup truck’s flat tire.He quickly types a keyboard command, causing an arrow on an electronic sign hanging over the righthand
lane to flash from green to yellow. The light warns drivers about the flat-fixers, hopefully allowing them to avoid an accident or maybe find a different route.
At first blush, the system at Mendoza’s disposal sure looks like a smart way to fight traffic. That’s what the federal government thought a decade ago when it began funding “intelligent highways,” a snappy term for a laborious program of installing sensors, video cameras and programmable signs along the nation’s highways. Today, systems in place in San Antonio and 49 other urban areas are indeed providing speedier accident response. But the cost is mounting, with the total taxpayer tab topping $8.5 billion to date. And despite that investment, controllers can’t detect traffic beyond where the sensors are installed. Worse, they have limited ways of alerting drivers; typically it’s either via signs or by notifying the news media. And as any
driver knows, even a 10-minute delay until the news breaks on the radio often means it’s too late to avoid the snarl.
In short, it’s time to hit the brakes on Uncle Sam’s approach to traffic management. Instead of a massive and costly new physical infrastructure that takes a couple decades to roll out, it turns out the road to truly intelligent highways is leading to cars themselves-or more precisely, to the wireless gadgets inside them. Tens of millions of vehicles are now loaded with cell phones, electronic toll-paying tags, onboard computers, two-way pagers and Global Positioning System receivers; all promise to play a role in a new era of wireless traffic management. That’s because, as millions of drivers gab on their mobile phones, the radio signals from those devices can double as handy traffic and speed sensors. Meanwhile, devices ranging from the lowly pager to luxury navigation systems are beginning to provide ways for drivers to get real-time traffic information,
customized for their routes. Even better: for-profit wireless companies seem willing to pick up part of the tab.
Plenty of curves and obstacles lurk along this new wireless superhighway. The technology needs refinement, business models remain unproven and drivers will want assurances their privacy won’t be invaded by new highway listening posts. But the promise is hard to deny. Around the country, from the crowded Washington Beltway to the San Francisco Bay, incipient tests of wireless traffic sensing are already hinting that these
new technologies can augment-and even eclipse-the original federal program, and do it in a matter of years, not decades. “With wireless technology, we don’t have to wait for the government to install loop detectors,” asserts Kenneth Orski, president of Urban Mobility, a Washington, DC-based transportation consulting firm. “Private enterprise can set up cellular networks faster and cheaper and extend intelligent highway capabilities to virtually every highway in the nation.”
Emissions And Enervation
New ideas for battling traffic can’t come fast enough. The Federal Highway Administration estimates that U.S. drivers spend 4.3 billion agonizing hours each year stuck on clogged roads. The average American now spends 36 hours per year stuck in traffic, up from 11 hours in 1982, according to a Texas Transportation Institute study released in May. The U.S. Environmental Protection Agency says tailpipe emissions are responsible for about 58 percent of U.S. emissions of carbon monoxide, 30 percent of nitrogen oxides and 27 percent of volatile organic compounds, among other pollutants. These bleak realities prompted Congress
to begin funding embedded sensor systems in 1991. But while the resulting “Intelligent Vehicle Highway Systems Program” did unleash technology against traffic, it was technology developed in the 1980s. Today, only about 10 percent of U.S. highways contain these sensors, a figure expected to increase to no more than 20 percent by 2020. Although the benefits of these systems vary by city, studies have generally shown they mostly help by speeding emergency response, with no hard proof that sensors have improved travel times.
All of which leaves the fast lane open for a wholly new approach to traffic management, one where radio waves replace magnetic loops as the key sensing technology. The clear leader of the wireless pack, experts say, is the cell phone. It was the mobile phone, after all, that first beat roadway sensors to the punch: drivers simply began calling 911 to report accidents they’d witnessed. Today, every American hillock, church steeple and high-rise seems to have sprouted a cellular radio transmitter and antenna.Americans own about 111 million cell phones-almost one for every two people-with 46,000 new subscribers every day, according to the Cellular Telecommunications & Internet Association, a trade group in Washington, DC.
Transforming this wireless communications infrastructure into a trafficsensing tool is the next step. Radio waves emitted by a driver’s cell phone during ordinary conversation can be used to pinpoint not only a car’s location but its speed and direction, too. The first application will be for emergency use: the federal government is requiring that, by October 2001, cell-phone companies be able to provide precise mobile-phone location when a 911 call is made. Eventually
these new location technologies could greatly extend and enhance today’s intelligent-highway infrastructure, too.
The leading approach to analyzing cell-phone signals to detect traffic patterns exploits the fact that these signals have distinct “fingerprints” that change as the phone’s location changes. That’s because a mobile-phone signal bounces off buildings, hills and other obstacles before converging on a cell tower, producing
a unique signal pattern for every spot along a roadway. Once these fingerprints are mapped and stored in a database, it’s possible to create software that
analyzes the signal of a passing cell phone-by sampling it several times per minute-to determine a car’s exact location, direction and speed.
Developing such a map is precisely the job of a white Ford E-150 van that regularly cruises a 30-kilometer stretch of the Washington, DC, Beltway, one of the nation’s epicenters of both traffic and talk (60 percent of area residents own cell phones). The van is owned by San Ramon, CA-based U.S.Wireless, a leader in the nascent business of generating traffic information. As the van tools down a congested stretch of I-495 from Springfield, VA, to Andrews Air Force Base, MD, a passenger talks on a cell phone. Every nuance of the signal fingerprints from that conversation is captured by a network of antennas and computers the company has installed on office and industrial rooftops lining the highway.
Then, thanks to a GPS receiver system in the van, each fingerprint is matched with an exact spot on the route. Later, when cars with cell-phone-chatting occupants drive by, the U.S.Wireless computer picks up the fingerprint, finds a match from its database, and-presto-spits out a location. By gauging how the fingerprint
changes over time, the system’s algorithm can calculate direction and speed, too. “Wherever we set up the network, we’ll be able to monitor vehicle density, speed and acceleration, and provide that information to anyone that’s willing to pay us,” says Howard Blank, U.S. Wireless’s vice president of technology.
To help support this grand experiment in cell-phone-signal cartography, the Maryland and Virginia transportation agencies are ponying up a combined $400,000. Initial results are encouraging, says David Lovell, assistant professor of transportation engineering at the University of Maryland, who is evaluating the test for the state of Maryland. The technology “tracks the trajectory of the vehicle continuously, which allows you to get a better feel for the pattern of congestion on the highway” than is provided by magnetic-loop sensors, he says. The test is continuing, but a report assessing its results is not due until December. Lovell says, however, that “everything appears to be working well so far.”
Still, not everyone is convinced by the tests. “It’s really a bizarre way to do this,” says Paul Najarian, director of telecommunications at the Intelligent Transportation Society of America, a research organization based in Washington, DC. “Every time a building comes up or goes down they have to recalibrateit. And the local topography changes according to the seasons. They basically have to run their van through the coverage areas over and over again to keep it all up to date.”
U.S. Wireless officials counter that recalibration costs are trivial compared to loop detector installation and maintenance. But while the company labors to perfect its technology, chief competitor TruePosition of King of Prussia, PA, is developing an alternative approach that never needs recalibration. TruePosition’s scheme is based on triangulation-determining cell-phone location from the times it takes signals to reach three or more stations. By also analyzing the angle at which a signal arrives, the company can accomplish this feat using just two towers, says Matthew Ward, TruePosition’s manager of strategic product development. Speed is calculated from changes in location over time, as with the U.S. Wireless technology. TruePosition is currently focusing on providing technology for 911 phone-finding. But Ward says the company plans to test its technology for traffic applications.
Beyond cell phones, there’s another wireless technology already beefing up traffic reports in some areas: those increasingly common toll-paying, windshield-mounted radio tags. Normally, a special tollbooth “tag reader” senses the passing device, records the code number associated with the owner’s account and subtracts the toll-and that’s it. But the E-ZPass tags used by more than four million New York-area drivers now double as speed and traffic detectors.
To use the tags for this new purpose, the Transcom coalition of regional transportation agencies mounts readers at regular intervals (ranging from 0.4 to 2.4 kilometers) along a highway. By analyzing the time it takes for a tagged car to pass between the readers, special software can calculate the speed of traffic along key arteries, with the results displayed in a regional traffic-management center in Jersey City, NJ. Just as Sam Mendoza does in San Antonio, operators publicize any snarls on electronic roadside signs or by alerting the news media. By the end of this year, more than 300 kilometers of highway stretching from Hartford, CT, to Trenton, NJ, will bristle with tag readers for speed detection, with Massachusetts and Pennsylvania eyeing the idea.
Ultimately, it might be possible to complement-or even replace-this growing arsenal of sensors with a third wireless device: the GPS receiver showing up in more and more cars, mainly as a navigation aid (see “The Commuter Computer,” TR June 2000). GPS could, in theory, provide a means of continuously tracking a vehicle’s location. But although traffic planners would love to collect all those rolling position figures and squeeze the numbers for speed and traffic data, it will be hard to pull off, at least for now. That’s because GPS receivers are just that-receivers, which determine position from incoming satellite signals-and don’t send position data unless a driver initiates a link, as when calling police for help or looking for directions. Such calls are relatively rare compared to cell-phone chatter; besides, reading a car’s position would involve eavesdropping on the content of a call rather than just sensing a signal.
Trafficking In Traffic
But before these emerging traffic-management technologies can really make their mark, some critical questions remain to be answered. One of the most basic is who will pay. “The technology is available,” says Najarian of the Intelligent Transportation Society. “But what’s missing is a revenue flow tying all these elements together.”
U.S.Wireless, for one, hopes to sell its data to state agencies, allowing them to broadcast updates via the standard news reports and warning signs, plus any future avenues that evolve. It’s also a good bet that many drivers will pay at least a modest fee to save some of those 4.3 billion hours spent stuck in traffic. Cue, a pager company in Irving, CA, is already charging $10 to $15 a month for personalized traffic information in more than 60 U.S. markets. Cue collects its data from all available sources, including loop detectors, helicopter news reports and, eventually, cell-phone signals. After a customer programs a route into a two-way pager, the company sends personalized alerts. These can be read as text on the pager or heard through a voice synthesizer.
Convincing people to pay for such services means getting them very precise information in a form they can use, notes Gerald Conover, a technology manager at Ford Motor and chairman of the International Affairs Council at the Intelligent Transportation Society of America. “Say I’m in Manhattan,” he says. “What I want is the real-time traffic environment on the street that I’m driving on, as well as the streets above and below me. I want the sensor data on a map so I can make instant decisions.”
On this front, GPS could really strut its stuff. While GPS won’t send traffic data, it might play a key role in helping drivers receive it. Using a wireless link, a driver might download real-time traffic data that an onboard computer could filter based on the car’s GPS-derived position. If the news was bad, navigation software could offer alternate routes. This information could be paid for by subscription, or delivered free along with advertising (geared to location in many cases: McDonald’s at the next exit!”).
But even if the right business models and communications methods are found, another concern remains: guaranteeing privacy. Drivers worry their toll tags, cell phones and other gadgets could be used to track their movements for market research, surveillance or to mete out speeding tickets, says James X. Dempsey,
deputy director of the Center for Democracy & Technology, a Washington, DC-based privacy organization. “The real concern here is that the information could be compiled and used to categorize, characterize and judge people,” says Dempsey.
The fears of E-ZPass owners on this front were tempered when Transcom agreed to scramble identifying information before launching the program in 1995. But the solution to the cell-phone privacy problem is somewhat less clear. Cell-phone trackers U.S. Wireless and TruePosition insist they don’t record caller identity on their networks. “We couldn’t care less who the callers are,” says U.S. Wireless’s Howard Blank. “We use dummy identifiers like caller number one’ and caller number two.’” But company promises aren’t backed up by law.While telecom companies are barred from disclosing someone’s identity without consent, Dempsey notes that upstarts like U.S.Wireless technically aren’t telecom carriers, and so they aren’t covered by such restrictions. The Cellular Communications Industry Association, a Washington, DC-based trade organization, recently proposed that the FCC develop privacy guidelines that include a provision for notifying customers of how their cell signals might be used.
Another view is that people will gladly cede some of their privacy for the chance to beat a traffic jam. Highway administrators in San Francisco, one of the first cities to explore cell-phone traffic sensing in detail, are about to put that view to the test. The region’s Metropolitan Transportation Commission is poised to sign a six-year, $5.2 million contract to study whether the U.S.Wireless approach can augment the brainpower of the city’s magnetic-loop-based “intelligent highways.”
“We’d like to use the system to develop highway speed profiles that drivers could use,” says Michael Berman, project manager with the Bay Area commission. First, though, they’re convening focus groups and conducting surveys on the privacy question.
Will the public accept it? That’s an open question. But has the technology arrived to lead traffic management into the wireless age? That call has already been made.
Become an MIT Technology Review Insider for in-depth analysis and unparalleled perspective.Subscribe today