Correlation Analysis for Exploring the Relationship Between Probe Vehicle Data and Event-Based Traffic Signal Performance Measures

Abstract

Probe vehicle data have been extensively used to assess the performance of roadway systems, particularly the interstate highway system. High-resolution data from traffic signal controllers is another data source used for automated traffic signal performance measures (ATSPM), which is increasingly used to evaluate signal operation. As high-resolution data require additional resources and may not always be available, probe vehicle data are sometimes used to evaluate signalized corridor operation. However, there has been little previous research comparing probe vehicle data on signalized corridors with ATSPM. In this study, we compared the average speeds from probe vehicle data with ATSPMs to examine the degree of correlation between the two datasets. Different scenarios including segments with random arrivals and platoons were considered for parts of US 20 in Dubuque, Iowa. Regression analysis was performed with average speed as the dependent variable to check the correlation between the two datasets. Four different signal performance measures, namely the percent on green, volume-to-capacity ratio, percent of green duration, and average delay, were used as independent variables. Two sets of categorical variables representing time-of-day and day-of-week variables were also added. It was found that there exists good correlation between the datasets, supporting the use of probe vehicle data for corridor-level analysis in the absence of high-resolution data. Additionally, the durations of the intervals used for data aggregation were varied to check its impact on the correlation. Higher levels of aggregation resulted in better correlation between the two datasets.