Quantitative analysis of probe data characteristics: Coverage,speed bias and congestion detection precision

Abstract

In recent years, there has been a growing desire for the use of probe vehicle technology for congestion detection and general infrastructure performance assessment. Unlike costly traditional data collection by loop detectors, wide area detection using probe-based traffic data is significantly different in terms of the nature of data collection, measurement technique, coverage, pricing, and so on. Although many researches have studied probe-based data, there remains critical questions such as data coverage and penetration over time, or the influential factors in the accuracy of probe data. This research studied probe-sourced data from INRIX, to profoundly explore some of these questions. First, to explore coverage and penetration, INRIX real-time data was illustrated temporally over the entire state of Iowa, demonstrating the growth in real-time data over a 4-year timespan. Furthermore, the availability of INRIX real-time and historical data based on type of road and time of day, were explored. Second, a comparison was made with Wavetronix smart sensors, commonly used sensors in traffic management, to explore INRIX’s speed data quality. A statistical analysis on the behavior of INRIX speed bias, identified some of the influential factors in defining the magnitude of speed bias. Finally, the accuracy and reliability of INRIX for congestion detection purposes was investigated based on the road segment characteristics and the congestion type. Overall, this work sheds light onto some of the less explored aspects of INRIX probe-based data to help traffic managers and decision makers in better understanding this source of data and any resultant analyses.