Using Intelligent Transportation Systems Travel-Time Data for Multimodal Analyses and System Monitoring

Abstract

Intelligent transportation systems (ITS) technologies and infrastructure are a potentially rich travel-time data source for travel-time mean and variance estimates. ITS data traditionally have been deployed and used in real time for passenger cars. How ITS data can be used for multimodal analyses and system monitoring is examined. The methodology used is applicable to any detector technology. Automatic vehicle identification (AVI) data were collected along a 3.2-km (2-mi) segment of US-290 in Houston, Texas. Simultaneous instrumented test vehicles collected travel-time data, and commercial-vehicle travel-time data were collected by video. The nonparametric loess statistical procedure was used to estimate the travel-time distribution properties as a function of time of day. The first application presented investigates how well link travel times from AVI replicate travel conditions for commercial vehicles. During congested conditions, average differences in travel-time estimates of 6.4 percent were found, whereas percent differences in coefficient of variation (reliability) were 14.7 percent. The research concludes that it may be reasonable to provide real-time traffic maps specifically for commercial vehicles. The second application investigates the accuracy of the AVI data for system monitoring. Estimated mean differences between AVI data and test vehicles were small (0.8 percent), whereas the ratio of the mean to the standard deviation (coefficient of variation) was relatively high (37.6 percent) during congested conditions. The AVI data source is found to provide a very cost-effective data collection method with which to estimate mean travel time while increasing confidence in the estimate.