Laurence R. Rilett
Date of this Version
The need for reliable performance measures of urban arterial roadways is increasing because of the rise in traffic congestion and the high value of travel time. Consequently, travel time reliability (TTR), which combines components of measures of central tendency and measures of dispersion of travel times, has recently received considerable research interest.
The basis of all TTR metrics is the travel time distribution (TTD). Estimating and forecasting arterial TTDs for TTR analysis is the focus of this dissertation. This dissertation proposes a new TTR methodology that is a marked improvement on recent TTR estimation and forecasting methodologies including the current US state of the art methodology which was published in the 6th edition of the Highway Capacity Manual (HCM6). The HCM6 TTR methodology is a very important step because it is the first reliability methodology proposed in the HCM. However, there is no evidence that the HCM6 TTR methodology has been calibrated with empirical TTD data.
The HCM6 TTR methodology was analyzed on four principal arterials in Nebraska. These corridors have historical empirical Bluetooth and INRIX TTD data. It was found that there were statistically significant differences between the HCM6 and empirical TTDs at a 5% significance level. More importantly, the HCM6 tends to severely overestimate the corridor’s reliability.
The sources and magnitude of the HCM6 error were investigated and a calibration methodology was proposed. It was shown that the calibrated HCM6 TTR methodology can replicate the empirical TTDs. Based on the preliminary work a new TTR estimation and prediction methodology was developed.
The contributions of this dissertation are threefold: (1) it provided the first comprehensive performance analysis of the HCM6 TTR methodology, (2) it developed a methodology for calibrating TTR methodologies, including that used in the HCM6, and (3) it developed a new TTR methodology that addresses the limitations of the HCM6 TTR methodology. Unlike the HCM6, the new TTR methodology can be used to estimate the population TTD and analyze changes in arterial roadway supply and demand components that impact travel time. Such changes may include the adoption of automated vehicles and the use of advanced signal controls.
Advisor: Laurence R. Rilett