Forecasting Traffic Flow

Intelligent transportation systems in the offing will ease traffic bottlenecks

6 min read

This is part of IEEE Spectrum's special report: Always On: Living in a Networked World.

The time is exactly 8 a.m. There are currently no delays on any of the city's major arteries, but the accu-traffic forecast calls for a major traffic jam causing a 20-minute delay within the next hour on the northbound..."

Just a minute!" would probably be the response if, say, a radio broadcaster were to claim the ability to predict what traffic would be like along a particular route an hour into the future. But this is exactly the ambition motivating the design of intelligent transportation systems. (ITS is an infrastructure being developed--along highways and city streets, and in cars and trucks--that can use information about traffic to speed up travel and make it safer.)

Several research groups (whose work is mostly funded by Federal and regional government agencies) have begun generating reliable information about what will occur on a road using predictive traffic modeling, as it is called. The status of a stretch of road--based on a number of variables measured by devices such as loop detectors embedded in the roadway and digital image acquisition and processing systems--is compared (using special algorithms) to a group of days when similar events occurred.

Groundbreaking work in this field is being done by the ITS Research Program at the University of Washington, Seattle. The group's Smart Trek project uses a tracking algorithm developed at the university to compare the positions of city buses to data generated during hundreds of previous trips. The entire Metro King County bus fleet (more than 1000 buses) has been outfitted with transponders that allow their position to be continuously monitored. The system continually updates a given bus's projected arrival time at the remaining stops along the route, and displays it at five of the area's busiest transit centers (where several routes intersect). This sort of information is another hallmark of ITS: providing the traveling public with a degree of comfort from knowing how long they may have to wait for the next bus or train. Bus location and progress are also displayed on a map in real time at the ITS program's Web site, https://busview.its.washington.edu.

Another project based on predictive traffic modeling is under way in Vancouver, B.C., along a 20-km route between Vancouver and the neighboring town of Richmond. As in Seattle, information about when buses will arrive at stops along the way (in the form of a countdown) is transmitted to electronic message signs at shelters along the route. But in this system, the buses can do something to prevent delays. By communicating via infrared transponders along the roadside, a bus can keep a green light from turning red as it approaches an intersection, or turn a red to green if there is no oncoming traffic in the cross street. Of 60 intersections along the route, 38 so far are equipped with signal preemption equipment. Limited service began in September; plans called for full service to commence last month.

These developments have not escaped the notice of the private sector. According to Daniel J. Dailey, electrical engineering professor at the University of Washington, and head of the university's ITS Research Program, several companies--like TrafficStation Inc., Los Angeles, InfoMove.com, Kirkland, Wash., and NextBus Information Systems Inc., Emeryville, Calif.--are hopeful that they will be able to turn a profit by providing traffic predictions, real-time mass transit information, or personalized traffic reports sent to a subscriber's mobile phone, pager, or personal digital assistant.

Of course, it remains to be seen how viable the market for such information is--especially in places like Seattle, where the University of Washington and the Washington State Department of Transportation provide the data for free. Another obstacle (that may be overcome with the introduction of next-generation wireless networks and devices) is that the small liquid-crystal displays of pagers and mobile phones may not be able to accommodate the detailed information (like the map of an entire city that gives an accurate spatial representation of the routes comprising the user's travel options) necessary for travelers to make decisions.

Mobile phones to aid traffic prediction and management

The Smart Travel Lab at the University of Virginia, Charlottesville, run by civil engineering professor Brian Smith, is among a number of sites looking for better ways to collect the traffic data necessary for predictive analysis. Said Smith: "Traditional [in-road devices like] loop detectors get the job [of counting cars and measuring their speed] done, but because they are buried in the pavement, installing and maintaining them is costly and time-consuming. We're looking at perfecting better methods of data collection, like cameras and image collection software, and tracking the location of cellular phones that many drivers have in their cars."

The tracking data would come from mobile phones equipped with global positioning system receivers or through triangulation from nearby cell base station towers. These two methods are both available but are limited--in the case of GPS, by the additional expense of adding receivers to handsets, and for triangulation, by the inability to tell the location of the phone within just a few meters. According to several ITS researchers--including Smith, Dailey, and Keenan Kitasaka of Translink (formerly the Greater Vancouver Transportation Authority, or GVTA)--tracking the position of those mobile phones gives a more detailed picture of what is happening on the road than do static devices like loop detectors or cameras. The difference, says Kitasaka, is that while loop detectors can provide the data used to determine traffic volume, headway between cars, and average speed across a particular spot, cell phones act as traffic probes that give the same information as well as the travel times for individual cars running through a municipality's highway system.

"[Tracking mobile phones] will revolutionize traffic monitoring, allowing us to realize the vision of homogeneous regional coverage of primary, secondary, and tertiary roads at a fraction of the cost of loop detectors," said Noah Rifkin of Veridian Technologies Inc., Buffalo, N.Y. The company, formed by the merger of Calspan Corp. and Veda Corp., offers vehicle-based services like an automated vehicle location (AVL) package that automatically dials into the local 911 system, providing a text message detailing the time of, say, a collision, and a report of its severity (based on the readings of three on-board accelerometers that set the system into action if any motion of the vehicle exceeds pre-programmed parameters).

Because mobile phone monitoring is such a rich data source, it is already changing researchers' understanding of how traffic incidents occur. More work is necessary, they say, but as this technique is tested and refined, the ability to forecast the state of a region's roads will be greatly enhanced.

A big test for mobile phone tracking

An important test bed for mobile phone tracking is Europe, where there are far more handsets, used more frequently too, than in the United States. The National Institute for Transport and Safety Research (Inrets) in Arcueil, France, SFR in Paris, and Autoroutes du sud de la France (ASF), Paris, have teamed up to carry out Europe's first test of a mobile phone tracking system. SFR is the subsidiary of French telecom carrier Cegetel that operates its Global System for Mobile Communications (GSM) network, and ASF builds and operates most of France's toll roads.

The 18-month project is the product of Serti (Southern European Road Telematics Implementation), the European Union committee responsible for the coordination of ITS implementation in 18 regions across five countries--France, Italy, Germany, Spain, and Switzerland. The test, begun last July, aims to identify, measure, and characterize the relationships between road incidents (such as collisions or stalled cars), traffic congestion, and mobile phone use--the initial hypothesis being that the number of people using their phones increases during periods of heavy traffic congestion, whether they are calling ahead to advise others of their status or simply to divert themselves.

The first part of the four-phase project--counting the number of phones in use along a 161-km stretch of an ASF-operated toll road south of Lyon--has been completed. According to Jean-Luc Ygnace, an Inrets researcher who recently completed a stint at California Path (Partners for Advanced Transit and Highways), the data revealed that roughly 80 percent of the cars had a mobile phone turned on as they moved along the route. Twenty-five percent of the cars used SFR's network--which, Ygnace said, is more than enough to generate the data necessary for establishing patterns.

Tracking mobile phones in cars will revolutionize traffic monitoring at a fraction of the cost of loop detectors

In the second phase, which, as of this writing, is still underway, Ygnace and his team at Inrets are analyzing the data to find any correlation between network usage and the level of traffic congestion. The third phase, slated to begin in February, will be the first time Europeans have attempted to use mobile phone locations for calculating travel times. This information will be compared with data obtained from loop detectors installed every 2 km along a 322-km roadway.

The project's final stage will entail mainly the evaluation of the technique. Said Ygnace: "If the technology works and the costs are reasonable, toll road operators and departments of transportation operating urban freeways might consider buying the traffic data from the cell carriers. In fact, on freeways that are not currently instrumented with loop [detectors], this solution can be a viable alternative." He also noted that the partnership is working on business models.

Mobile phone tracking gets government backing

Promising as this technique seems, there are significant barriers to its widespread use. So far, drivers with cell phones have signed on voluntarily. But tracking handsets is a serious privacy issue that is being debated. Governments and consumer groups want assurances that information about callers' whereabouts--which can be tracked any time their mobile phones are turned on--will never be abused, but those who have signed on feel the eventual benefits are worth the risk.

Mobile phone tracking was given a big boost in the United States by the Wireless Communications and Public Safety Act of 1999 [911 Act], which designated 911 as the universal emergency number and mandated that wireless carriers provide emergency dispatchers with the location of a handset from which a 911 call is placed. By the end of 2001, 25 percent of new handsets must be able to relay their location according to Federal automatic location identification (ALI) standards, half must meet the standard by June 2002, and all new handsets produced from 2003 on must be ALI equipped.

To Probe Further

The Intelligent Transportation Society of America's information clearinghouse is at https://www.itsa.org.

Go to introduction

This article is for IEEE members only. Join IEEE to access our full archive.

Join the world’s largest professional organization devoted to engineering and applied sciences and get access to all of Spectrum’s articles, podcasts, and special reports. Learn more →

If you're already an IEEE member, please sign in to continue reading.

Membership includes:

  • Get unlimited access to IEEE Spectrum content
  • Follow your favorite topics to create a personalized feed of IEEE Spectrum content
  • Save Spectrum articles to read later
  • Network with other technology professionals
  • Establish a professional profile
  • Create a group to share and collaborate on projects
  • Discover IEEE events and activities
  • Join and participate in discussions