Modeling the Deterioration of Inductive Loops

Abstract

Inductive loop detectors have been widely used in transportation system, particularly to monitor the performance of highway systems such as traffic flow, speed, and vehicle classification. Accurate data are dependent on the quality of the detector mechanism and the level of maintenance of these devices. Yet, this aspect of maintenance has been neglected both in the field and in research studies. A review of the lifetime data of loops revealed the following features: (1) lifetime data for a specified loop was not available from existing maintenance records; (2) lifetime data were censored; and (3) the lifetimes of loops located in different stations were dissimilar. For these reasons, a censored random-effect panel data duration model was used in this study that can deal with heterogeneity and censored duration data. A formula was developed to estimate the lifetimes for individual loops, despite the unavailability of lifetime data in the original maintenance database. Because vehicles of different classifications were distributed differently in adjoining right lanes, different models were adopted for two different traveling lanes. The models indicate that the lifetimes of loops are highly determined by the volumes of truck traffic; a finding that can be used to assist loop detection maintenance scheduling.