University of Minnesota Driven to Discover
U of MNUniversity of Minnesota
Center for Transportation Studies

Evaluating GPS for Assessing Road User Charges

Presentation by Dr. Pi-Ming Cheng, Dept. of Mechanical Engineering

September 9, 2003

Today, vehicle fuel taxes are the primary means of assessing charges to road users to cover the costs of constructing and maintaining roads. Changes in vehicle technology—such as the increasing popularity of electric vehicles and the impending introduction of vehicles powered by fuel cells—are prompting researchers to seek new mechanisms to recover road user fees. One such mechanism, currently under study at the University of Minnesota and the University of Iowa, is based on Global Positioning System technology.

In principle, the system is straightforward. A GPS receiver on each vehicle determines the precise location of that vehicle, and compares it to a detailed digital map. This allows an onboard computer to determine on which road, and in which jurisdiction, the vehicle is traveling. Data on road use and mileage can then be transferred to base stations, and a fee statement forwarded to the vehicle's owner.

In practice, numerous engineering challenges must be overcome before such a system can be deployed. Pi-Ming Cheng of the Department of Mechanical Engineering, whose work on the project involves evaluating the suitability of GPS components, presented an overview of issues related to GPS and digital mapping for road user charges assessment at a seminar September 9, 2003.

Cheng's presentation focused on establishing the accuracy and reliability of the GPS subsystem. Because position finding is accomplished by a continuous series of measurements of varying precision, this accuracy has to be expressed statistically—as spatial precision at a specified error rate. Current consumer GPS receivers have a measured accuracy of roughly ±13 meters at 95% certainty. In order to reliably differentiate between roads in close proximity—such as frontage roads and main highways—accuracy of ± 6 meters laterally at more than 99% accuracy is required.

Because consumer-grade GPS is not up to the task, Cheng tested a variety of professional-grade differentially corrected GPS (DGPS) options. These systems boost the accuracy of receivers by providing a correction signal that helps the receiver determine its position more accurately than would be possible using the standard system alone. Correction signals can be broadcast from ground stations or from private satellites.

The GPS systems tested ranged from consumer-grade to esoteric high-end systems designed for critical navigation applications. Components were judged against a "gold standard" professional receiver boasting centimeter-scale accuracy and a price of over $20,000—far beyond the reach of most automobile buyers.
All the GPS systems were mounted on a test vehicle and driven on a course that had been digitally mapped. Longitudinal and lateral deviations were recorded for each system, and the degree of error calculated by comparing the test system's reported location to the location reported by the reference system.

Cheng's findings indicate that at least some road-user charge implementations could be possible with currently available GPS receivers; however, the cost of GPS equipment is a significant factor in developing a product that would be accepted by auto buyers.

Additional factors evaluated in the test included the effects of tall buildings and hilly terrain on the GPS systems' accuracy. The "urban canyon" effect of heavily built-up downtown areas was found to be a significant problem for the GPS receivers, because signals from satellites and correction stations were frequently interrupted by tall buildings. Time required to recover the signal after losing it was a problem for all receivers. Hilly terrain, on the other hand, proved to be no great difficulty for the systems tested, as they maintained their accuracy rates in spite of elevation changes.