Application of Maximum Entropy Sampling Design to Traveler Information System Data-Quality Evaluations

Abstract

Evaluation of travel-time data quality from traveler information systems (TISs) is a critical need for many transportation agencies. To measure data quality, observations of ground-truth travel time must be collected on links in a transportation network. Therefore, selecting which links in the network to monitor is a critical step in evaluating the quality of a TIS. This paper presents a maximum entropy sampling (MES) design that can be used to objectively select a subset of links from a road network for ground-truth data collection. This approach objectively identifies subsets of links that have the greatest variability in traffic conditions. This allows for a more rigorous evaluation of a TIS to determine the system’s ability to effectively