The authors would also like to acknowledge the support of NITC (National Institute for Transportation and Communities) Transportation Center for funding this research effort.
Transportation Research Record: Journal of the Transportation Research Board
Bus lines, Travel time (Traffic engineering), Global Positioning System -- Data processing -- Evaluation
The recent availability of high-frequency transit (HFT) GPS data for buses has allowed the estimation of detailed bus-travel speed profiles between bus stops. HFT data are defined as data comprising GPS vehicle trajectory recorded at or less than 5 s intervals. With HFT data it is now possible to measure changes in bus speed at specific locations of interest, such as intersections, ramps, crosswalks, and so on. Previous research efforts have not compared GPS-based bus speeds with general traffic speeds at specific locations between stops. This research fills this knowledge gap, utilizing accurate stationary radar speed data as a baseline or ground truth data to estimate GPS-based bus speed accuracy. A thorough data analysis of the bus and traffic speed data indicates that HFT speed estimations between stops are accurate and highly correlated with traffic speeds. Time–space speed profiles and regression analysis are utilized to quantify factors that affect HFT speed estimation accuracy. The relative advantages and limitations of the HFT data are presented and discussed. This study concludes that large HFT datasets can be utilized to accurately monitor speeds between transit stops. HFT datasets are suitable to cost-effectively monitor recurrent arterial speed performance for both passenger and transit vehicles.
There are several well-established technologies for collecting speed and travel time data, including vehicle identification sensors—like Bluetooth readers, radar devices, loop detectors, and probe car data. Loop detectors and vehicle identification sensors, once installed, continuously record most (or a share of in the case of Bluetooth readers) vehicles passing specific road sections. Bluetooth readers sample a fraction of the traffic and can estimate average speeds between two readers, but cannot produce speed profiles between the reader locations. The density of Bluetooth, radar, or loop detectors is typically very low on most arterials and most non-freeway network links do not have speed profile data.
This research utilizes high-frequency transit (HFT) bus (probe) vehicle data to estimate arterial street speed and produce speed profiles along a designated arterial. The GPS data are denoted as high frequency when GPS points are obtained at or less than 5 s intervals; the vast majority of transit agencies in the USA are currently storing 30 s interval GPS data or not storing any GPS data outside transit stops. Dedicated probe vehicles can be utilized to collect travel time and other data for designated routes in the network. However, because of cost considerations, government agencies cannot run probe vehicles continuously; moreover, privacy concerns make it problematic to gather data from private vehicles. Private data vendors can provide crowdsourced probe vehicle data, but these datasets can be expensive, and in many cases the data are already aggregated or preprocessed.
The recent availability of HFT data for buses has allowed the estimation of bus-travel speed profiles between bus stops. HFT data are defined as data comprising GPS vehicle trajectory records at or less than 5 s intervals. With HFT data it is now possible to measure changes in bus speed at specific locations such as intersections, ramps, crosswalks, and so on. No previous research efforts have compared GPS-based bus speeds with general traffic speeds at specific locations between stops. This research fills this knowledge gap. This research utilizes accurate stationary-radar speed data as a baseline or ground truth data to estimate GPS-based bus speed accuracy. More specifically, this research is original because it aims to answer two novel research questions: (1) how accurate are the speed measurements obtained utilizing HFT data when compared to state-of-the-art stationary speed sensors; and (2) what are the factors likely to affect the accuracy of HFT speed estimations at specific locations between bus stops?
To analyze the accuracy of the HFT speed data, exploratory data analysis, correlations, time–space speed profiles, and regression analysis are utilized. The final sections of this study analyze the impact of transit vehicle frequency and accuracy, as well as advantages and limitations to the HFT speed data. The next section provides a brief literature review.
Locate the Document
Figliozzi, M. A., & Stoll, N. B. (2018). A Study of Bus High-Resolution GPS Speed Data Accuracy. Transportation Research Record, 2672(8), 187-198. https://doi.org/10.1177/0361198118793273