Real-time traffic data holds the key to minimizing congestion's impact
Stay or go? Grind it out at 5 mph and hope the traffic clears soon, or jump off the exit and hope the peripheral roads are clear? It's a gamble in an economy that leaves fleets no margin for wasted time or unnecessary miles, a bet that's being waged every day throughout the country as traffic congestion ties up more and more of our roads.
Up to this point, you've had to rely on traffic's version of The Daily Racing Form — historic records that can identify congestion trends for Interstates and other main highways, supplemented by traffic reporters handicapping routes and alternatives based on limited eyewitness information. If you're dealing with a recurring congestion pattern, or if the on-site traffic report isn't outdated and if there's reliable, timely information on the possible alternate routes, you'll be able to make an informed decision. But that's a lot of ifs, and most of the time you or your drivers are going to make a best-guess decision and hope it's the right one.
The solution is fairly easy to identify — real-time traffic data that could be used to quickly evaluate the congestion right in front of your driver's windshield and determine whether there are viable alternatives. From the first day a traffic reporter took off in a helicopter, we've been on our way to real-time traffic service, but it's proved to be much more complex than simply putting eyeballs in the sky or cameras on the side of the highways, especially when you need reliable information for help making nearly instant decisions for dozens, if not hundreds, of trucks on a daily basis.
Depending on who you talk to and how they define real-time traffic data, we finally may have the resources fleets need to minimize congestion productivity losses on a case-by-case basis. At the very least, we are on the verge of widespread availability of information on traffic that's accurate, that covers broad areas and that can power dynamic routing around problems. Deciding if it has value for your fleet requires understanding what goes into creating real-time traffic data, what different providers mean when they offer real-time information, and what it can do when combined with routing and navigation systems.
Today, the most common real-time traffic information comes from radio, TV and web reporting. “It's really pretty good for what it does, which is give a broad brush picture of what's happening at big targets like bridges,” says Dr. Alain Kornhauser, founder of the mapping and routing software provider ALK Technologies, as well as professor of operations research and financial engineering at Princeton University. “But that doesn't tell you what to do for the trip you're taking, especially if it's not in a targeted area. It's really ‘traffic-tainment’ that doesn't help with the millions of other trips drivers are making. … It assumes things are okay when there's no knowledge of what's happening” outside those monitored points.
Fleets aren't interested in traffic as entertainment, but rather want to know which way they should route their trucks. “There may be thousands of options [for any particular route], so you need a way to rank them for the best solutions at any given time, and a computer is perfect for that,” says Kornhauser. “It can determine the best route and tell me what it is. I can't do that, and you can't ask a driver to make those decisions. To do that, you need data, pretty good data, not just on the highways but on all the other roads around the highways that could be alternatives. And it needs to be mechanized so you can put it into [navigation] algorithms that can give the driver turn-by-turn directions to get around the problem.”
ALK has identified one million road segments that cover everything but neighborhood streets in the U.S., and collecting good real-time or near real-time data for so many segments would seem like an insurmountable obstacle to developing a dependable traffic system for fleet operations. But Kornhauser believes we already have the technology in place to do just that. “We have many vehicles equipped with GPS systems that know where they are and how fast they're going,” he says. “And with cellphones, we have the communications to send and receive that data.
“In computer architecture, that's a million ‘buckets’ and whatever the probes are experiencing can accumulate in those buckets, so we have individual entities to work with on a computational level,” Kornhauser says. “You just send the information from the buckets around [the vehicle], which is not a lot of data, and let the onboard device compute the better route with directions.”
ALK has already shown with a limited demonstration project that this approach for “anywhere to anywhere” real-time traffic data can work, but one obstacle remains: Who collects and organizes all that data to fill the million buckets?
ONE OR MANY
“There are probably a few carriers that are large enough, that have enough probes in their fleet to populate [the data buckets] in the places they serve,” Kornhauser says. “All they need is the computational architecture to run the service on their own servers. Maybe a consortium of carriers could do it, or a community of local fleets in a metro area. And since they're truck operations, you'd also have their operational characteristics — top speed, grade speed and so on, which are different than for cars.”
The cost of developing and maintaining such a real-time system “is infinitesimal when I compare it to the operational costs of a fleet,” says Kornhauser. Using real-time traffic to navigate through congestion “might not be useful every day, but there are going to be times when it will save a driver an hour or two, and it doesn't have to do that often to be valuable. It can help the fleet, their drivers and their customers. That's why I'm so passionate about it.”
As part of a four-prong approach to providing traffic services, INRIX offers four types of information to a variety of companies like Tele Atlas, TomTom, MapQuest, TeleNav and other suppliers of navigation and routing services. While it began by building historical average speeds for approximately one million miles of roads in North America, it has since added real-time data on traffic flow and incidents that affect that flow for all Interstate highways and some arterial roads. The fourth service, called Total Fusion, combines data from all three sources.
While still far from Kornhauser's goal of a million road segments, just last month INRIX announced that it has tripled its real-time traffic coverage across North America from 55,000 to 160,000 mi. and 126 metro areas with over one million commercial vehicles and cars delivering real-time GPS data to the company.
“Where historical traffic flow is shown to be accurate more than 85% of the time during normal congestion periods, real-time flow helps identify periods of heightened congestion due to accidents or construction or incidents,” according to an INRIX study on traffic data quality. “From a routing perspective, real-time flow and fusion flow data drastically improve routing accuracy and anticipated travel time.”
The report goes on to point out that “the significant improvements in the quality of traffic data reporting over the past several years are now providing higher value to consumers, commercial businesses and the public sector. Finally, real-time, historical and traffic fusion information can be trusted for routing, current and expected traffic flow, and transportation planning decision-making.”
“We work with a number of mobile resource management companies in the trucking industry to aggregate traffic data from their onboard devices,” says Scott Sedlik, INRIX vp-marketing. While those devices might have initially reported vehicle operating data on a daily basis, INRIX now gets reports as frequently as every minute, providing latitude/longitude positions, vehicle headings, ignition status and even vehicle type.
“We have truckload and LTL trucks, utility trucks, in-city delivery vehicles, taxis and even concrete trucks,” says Sedlik. “For every source, we classify the vehicle since they have different driving patterns, and we need to account for that in the analytics. We are also getting consumer [automobile] data and can mix in that information intelligently to determine speed and give us richer information, especially on arterial roads.”
“Three years ago, there were issues across the board with accuracy and quality [of real-time data], but it's radically different now. We're getting 1.5 billion GPS data point reports a month now,” says Sedlik. “Not only have the number of vehicles and the number of vehicle reports increased exponentially, but we've put systems in place that help us proactively go after probes in areas where we need data. We manage it from a portfolio perspective of sources to ensure we can report accurately.”
Currently, INRIX has real-time speed information for “all freeways and highways, which are most of the roads used by long-haul trucks” Sedlik says. “Our next focus is major arterials, but they're tricky. Take stoplights, for example. What speed do you report [for an arterial] when vehicles are stopping for lights?” A new agreement withMotor Co. to add INRIX real-time traffic and routing to its Sync onboard electronics systems next month should expand the reporting coverage of the systems.
Delivery of INRIX real-time information to fleets and their drivers can take a number of routes ranging from direct subscriptions to third-party service options; however, another alternate now being explored is direct distribution by government or public highway groups.
Last summer, the I-95 Corridor Coalition, a partnership of transportation agencies, toll authorities and other organizations covering 16 Eastern states, launched the first public-sector program to seamlessly cross state borders to provide real-time traffic information to drivers, including areas where no data previously existed. Under contract to the coalition, INRIX initially provided coverage for 100 mi. of I-95 stretching from New Jersey to Virginia. Judged a success by the group, the project received an additional $6.4 million from the U.S. Dept. of Transportation to expand testing. The next phase calls for expanding coverage to about 1,500 mi. of freeways and 1,000 mi. of major arterials from New Jersey to North Carolina. Eventually, it could extend the entire length of I-95 from Maine to Florida, according to INRIX.
While onboard routing around congestion is the most obvious application for real-time traffic data, Sedlik points out that it has important implications for fleet back-office operations as well. While fleets have used the company's historical traffic information for routing for some time, some fleet customers will begin integrating real-time data into their fleet management systems by the second half of this year, according to Sedlik. Possible applications range from giving dispatchers and customer service representatives more accurate pictures of transit times to improving fuel optimization applications and hours-of-service management.
“I'm especially excited about the explosion in mobile [communications] and connected navigation devices,” says Sedlik. “This is the next evolution [in traffic services] because it allows us to deliver much more personalized information into the truck cab.”
For most fleets, just reliably helping a driver navigate around productivity-robbing congestion will be exciting enough.