For decades, traffic conditions have been measured using aggregated pointmeasurements from loop detectors. However, new technologies have become recently available that can assess traffic conditions by tracking vehicle trajectories and travel times. Among these new technologies is cell phonetracking, a concept that has received strong curiosity from both public and private entities in the transportation community. A private firm in the United States has constructed a system that, using the Sprint PCS network, can track cell phone movements in Minneapolis, Minnesota, and deliver travel times for most of the urban roads, including both limited-access freeways and signalized arterials. A University of Minnesota research team was consulted to evaluate system's travel times against ground truth conditions and then assess the level of accuracy and reliability of the technology through statistical analysis. The purpose of this paper is to illustrate the system's accuracy, in terms of travel time and speed data accuracy when compared to real conditions, during its first-stage deployment in Mayand early June of 2007. This paper discusses the cell phone tracking system's performance on a limited-access freeway and a signalized arterial, starting with the data collection strategy used for each road and leading to the final result. It concludes by saying that, with the small sample ofrepresentative data in the paper, this technology produces results with varied accuracies. Whether or not this system would produce acceptable margins of error for speeds and travel times is dependent on the guidelines set forth by interested transportation agencies.
Abstract