Resumen
The objective of this paper is to develop a methodology for forecasting freeway vehicle travel time reliability for transportation planning using probe GPS data. Travel time reliability is measured using the coefficient of variation of the GPS spot (instantaneous) speed distribution. The proposed approach establishes relationships between travel time reliability and roadway traffic density in order to forecast reliability given future traffic conditions. The travel time reliability and traffic density datasets are segmented into different homogenous groups using the K-means cluster algorithm and the corresponding reliability-density relationship of each cluster is fitted by minimizing squared errors. This paper employs a truck probe GPS dataset as an example to demonstrate the proposed approach. The approach can be applied with any GPS datasets for forecasting reliability.