Using Real-Life Dual-Loop Detector Data to Develop New Methodology for Estimating Freeway Travel Time Reliability

Abstract

Travel time reliability is an important system performance measure for freeway traffic operations. It captures the variability experienced by individual travelers, and it is an indicator of the operational consistency of a facility over an extended period. Real-life transportation data were used to develop a new methodology for estimating travel time reliability of the I-4 corridor in Orlando, Florida. Four different travel time distributions were tested: Weibull, exponential, lognormal, and normal. The developed best-fit statistical distribution (lognormal) can be used to compute and predict travel time reliability of freeway corridors and report this information in real time to the public through traffic management centers. When compared with existing Florida and buffer time methods, the new reliability method showed higher sensitivity to geographical locations, which reflects the level of congestion and bottlenecks. Another advantage of the new method is its ability to estimate the travel time reliability as a function of departure time.