There is No Silver Bullet - Accurate Traffic Information Requires Multiple Data Sources
Nothing frustrates a commuter more than hitting traffic
congestion on the way to work. And nothing's more maddening to a
commercial truck driver than seeing the navigation system report normal
traffic on the interstate while he's stalled and wasting
$4-plus-per-gallon diesel fuel for a half-hour as police clear a
traffic accident. Such are the tribulations triggered by poor traffic
Despite progress this decade in providing more up-to-date traffic
information to auto and commercial truck drivers - whether via a Web
portal, a Global Positioning System (GPS) device, a mobile phone, a
tech-savvy radio or TV traffic advisory service - vast improvements are
still needed to help forewarn drivers to find alternate routes when a
traffic accident snarls the freeway.
The data provided via these services haven't yet reached a level of
quality on which drivers can fully count to deliver accurate travel
information. Today, individual solutions rely too heavily on a single
or very limited number of data sources to produce real-time
information. It's an approach that may result in substantial data gaps
and inaccurate information. But real value will emerge when these
various data sources are dynamically analyzed and proactively monitored.
One key indicator underscores the necessity for improved real-time
traffic data: commuting time. In the U.S., for instance, the average
commuting time has climbed to 25.1 minutes in 2005, from 22.4 minutes
in 1990 and 21.7 minutes in 1980, according to the U.S.
periodic report on commuting, 2005.
In major cities, of course, those average times are much higher,
including 34.2 minutes one-way in New York.
But U.S. commuters don't have it the worst. In Canada, the average
commute time has increased by nine minutes since 1992, with commuters
now spending 63 minutes getting to and from work in cities across the
country. Toronto tops the list, with the average round trip taking 79
minutes, according to Statistics Canada.
What's the cost to commuters of those longer driving times? Consider
that a 10 percent reduction in an hour-long commute - just six minutes
per day - would provide an entire weekend's worth of free time per
year. Further, that extra 30 to 50 hours of driving time a year, per
person results in higher fuel consumption, lost productivity at work
and poorer air quality.
Longer commutes interfere with living and working conditions, raising
stress levels. Less time is available for discretionary leisure
activities with the family. Long-term effects on health must still be
investigated but it does contribute to increased absenteeism at
work. And then there's the stress of paying as much as $500 a
year, maybe more, at the pump.
What's the Problem?
So, why aren't updated traffic data getting to commuters? While road
sensors have become more sophisticated - involving microwave radar,
infrared sensors, ultrasonic detectors and passive acoustic devices
that can be attached to bridges, overpasses and lighting structures -
such roadway sensors and travel-speed detectors cover less than seven
percent of all major roads in the U.S. Some individual urban areas have
zero coverage. Those percentages are much different in many other
That coverage is simply not sufficient to provide the real-time traffic
information that drivers, commercial and otherwise, require today. And
since the information trend this decade has centered around collecting
speed information on the road and travel time calculation, the
inadequate data hurt performance.
But it's more complicated than that. Even with advances in GPS and
other probe data collection, speed calculation is not the only measure
of advanced traffic information. At the most extreme, when a highway or
road must close for a bridge collapse or repair, police stop traffic
for miles before that bridge. The road sensors show no cars traveling
so that suggests traffic is flowing freely. A more sophisticated
modeling engine is necessary to determine when to reduce the weight
that traffic models give sensors or GPS flow data. Here's another
complication. Chicago employs an excellent road sensor system, but when
snow is falling at a rapid two inches an hour, the sensors don't
provide a good estimate of travel times. They're only one source of
In addition, solid reports of incidents are still essential. Collecting
speed data can pinpoint major traffic slowdowns. But unless the cause
of the slowdown is determined - a major hazardous materials spill or
simply a flock of chickens getting loose - the impact model will be
imprecise, yielding a less meaningful travel time forecast and
uncertain advisories to drivers.
Wanted: Multiple Data Sources
Traffic and travel experts increasingly recognize the need to use
multiple sources of real-time traffic data to develop dynamic and
meaningful travel advisories. They're realizing that no silver bullet
is going to emerge to provide all the real-time data essential to
delivering reliable traffic information. Still, technology advances are
arriving which, when coupled with other means of compiling traffic
data, are improving substantially the flow of travel advisories and
information that help drivers.
Already, a number of companies are moving swiftly into the broad area
of floating vehicle data to calculate road speeds. In the U.S., cell
probes, GPS probes and eventually tracking data from cell phones are
all being researched and evaluated. Key variables include data sampling
requirements, latency and "denoising" of extraneous data samples.
In Shanghai, cell phones are being used as part of what is considered
the world's largest cellular-probe system to develop accurate real-time
traffic information. The anonymous mobile phone position and signaling
data in China Mobile's GSM network are collected and converted into
travel time and speed information for hundreds of miles of major
highways and surface streets in the city. Millions of Shanghai
subscribers will soon receive reliable traffic information to help them
deal with that city's major traffic congestion problem. The
state-of-the-art cell-probe deployment is helping support logistics
management for the 2008 Beijing Olympic Games and the 2010 World Expo
No Silver Bullet
As more companies go global, they increasingly want to obtain reliable
real-time traffic information in many locations. They need that
information to ensure that time-critical packages are delivered. For
instance, a growing number of companies want accurate traffic
information to help them manage their supply chains and fulfillment.
The accuracy of traffic information today is probably where weather
forecasting was in 1990, so there's plenty of catching up to do before
extremely reliable traffic and travel data will be available. Still,
the gap is closing fast, and traffic science will accelerate the
process. But a single silver bullet won't provide the answer. And
unless we take a new approach to end the reliance on a silver bullet
data source, the data gap will persist.
The real remedy will involve using myriad sources of traffic and
weather data - some already available, some being tested now and some
still to be developed - and combining them to forge one dynamic source
of real-time traffic information. Then the promise of Dynamic Traffic
Assignment, the ability to more accurately predict traffic flow across
the entire road network and intelligently advise drivers of alternate
routes, can become a reality.
And when that time arrives, it will give all drivers peace of mind.
Published Thursday, July 24th, 2008