GPS Analysis Yields Real-Time Forecasts

By Bruce Lilly, Contributing Writer

This story appears in the June/July 2013 issue of iTECH, published in the June 10 print edition of Transport Topics. Click here to subscribe today.

Advances in traffic avoidance routing technology are made possible by two things: the vast amounts of data being collected and the complex algorithms that transform the growing mountains of raw data into useable information.

Two of the leading companies in this area are HERE, formerly known as NAVTEQ, and INRIX. Nokia acquired NAVTEQ in 2008 and rebranded the business unit as HERE in 2012.

A huge part of the data is collected by tapping into anonymous GPS data from fleets and consumer mobile devices. When you drive, you may be helping establish historical traffic records.

Mike Finn, HERE director of traffic for Americas, explained how the company collects GPS data. “Our customer base works with us in a virtuous circle,” he said. “They send us data they collect from all of the GPS units in the fleets of their customers. We also collect data from millions, if not tens of millions, of handheld devices that have GPS data.”

HERE collects these GPS data, which are called probe data, assembles them to its map and runs them through its model’s algorithms. Then the digested data are sent back to HERE’s customers, completing the circle.

“We actually reproduce very closely with very little delay what the prevailing speeds are” on highways, Finn said. “We collect billions and billions of these points live to tell you what’s happening live. We also use this data to create our historical databases.”

The ability to build these databases is relatively new. “It’s taken a few years for this traffic technology to mature,” Finn said, “but we are now getting a sufficient volume of data inputs to put out a highly reliable product. We’re seeing this grow in a massive way, much like navigation did 10 years ago.”

INRIX’s sources for GPS data closely match HERE’s sources and include vehicles equipped with Ford Sync or Toyota Entune, the two-way connected navigation systems that are built into many new cars. The collected data are used to paint a traffic picture not only of highways and interstates but also of city streets and arterial roads.

“What we love about consumer data is that consumers drive where the fleets don’t go,” said Jim Bak, INRIX director of community relations. “They go deep into the neighborhoods, so it allows us to get a more complete picture as we go deeper into the road network.”

In addition to probe data, HERE and INRIX collect sensor data from state transportation departments and transit authorities. The distinction between probe data and sensor data is that the probe data coming from GPS units show what individuals are doing everywhere, while the data from sensors show what every- one is doing at specific locations along highly traveled routes.

“There are sensors along various roads in large metropolitan areas throughout the United States that collect data by lane and track how fast everyone is going,” Finn said. “HERE also has its own sensor networks in 27 markets throughout the U.S. and Canada.”

All of this automation is supplemented by contact with other humans. “We interact directly with people within local agencies,” Finn said. “This gives us access to information about road closings and restrictions. HERE personnel also work with radio responders and scanners.” Bak stressed that the wrong conclusions could be drawn if the data aren’t handled the right way. “If you don’t have good analytics, you’re going to misread taxis that are stopping to pick up fares and delivery trucks that are stopping to make deliveries,” he said. “You’ll see them as cars in traffic jams. We have to run some very smart analytics on top of this data to understand where these outliers are.”

Another feature offered by INRIX is access to images from traffic cameras. Those images can be sent to mobile phones, displayed on a Web page or viewed on a connected navigation device.

“People may see a map that says traffic is jammed along a certain road, but maybe they’re skeptical, so they want to see an actual photo to be sure,” Bak said. “The traffic cameras offer the opportunity for verification. Truck drivers like to be able to see what the road conditions look like over mountain passes in the winter.” Bak also suggested that traffic data can help carriers find new revenue sources. “This traffic data allows fleets to set pricing in accordance with actual driving times,” he said. “If a customer needs a rush delivery to a location that will involve heavy traffic, the carrier can scale pricing to reflect the extra time and fuel needed to make that delivery. This makes it possible to drive more revenue by adjusting the prices as needed.”

Finn and Bak see new applications arising in the future and suggested that cars will be providing important data. One example is data from traction control and anti-lock braking systems. Data could come in showing that the traction-control systems were activated for several cars in the same location within a few minutes of each other. The map shows that the location is a bridge over a river, and the weather data indicate near-freezing temperatures at that location.

“The system will pull this data together and know to issue a warning that the bridge is freezing and should be avoided,” Finn said. Knowledge of this sort is highly valuable.

“This also will help transportation agencies determine where to send trucks with salt to deice the roads,” Bak said. “There’s a lot of data in the cars. We’re not yet getting that data in real time, but when we do, there are things that we can do with it.”