Characteristics of Cell Phone Probe Technologies and Field Testing of Very High Volume Probe System

Numerous studies of cell phone based probe systems have been conducted over the past 10 years. The study results have varied widely leading to uncertainty as to how accurate cell phone based traffic systems can be and what degree of network coverage can be achieved. Much of the variability in study results comes from comparing very different technologies and implementations. For the traffic practitioner, and even the researcher, cell phone probe systems are typically black boxes with little or no implementation details available. This hides the significant differences in both quantity and quality of input information available from a cellular carrier. This paper examines the most common of the technologies used in cell phone probe systems and provides an empirical comparison of the data streams available. Depending on what and where data is pulled from the carrier, the accuracy of individual locations can vary by an order of magnitude and the volume of locations can vary by 2 orders of magnitude. An examination of the accuracy of a system using Abis measurement reports, the most voluminous cell phone data available, was undertaken to show the potential of the very high volume location data. The examined system, when compared with loop detector data, produced traffic speeds within an average of 4.6 kph (2.8 mph) of loop speeds, a 2.9% difference, and within an average absolute difference of 8.6 kph, an 11.6% difference.