An Investigation of the Application of Stratified Sampling in Probe-Based Traffic-Monitoring Systems

Abstract

Probe vehicles represent samples in a traffic stream, and hence, probe-based traffic monitoring relies on effective sampling of vehicles. Current research views sampling from a networkwide perspective, and a single sample size is specified for the entire network. Using this method, research has indicated that the smaller routes in a network, such as minor and other arterials, are not adequately monitored. Increasing the sample size is not always a feasible option because of the restrictions of cost for use of the wireless network bandwidth and the availability of required number of probe vehicles. In this article, the authors describe a research effort that identified and extended stratified sampling as a method for increasing the efficacy of probe-based traffic monitoring. A methodology for application of stratified sampling for probe-based traffic monitoring is developed and implemented using a heavily traveled suburban traffic network simulation as a case study. Also, the authors discuss results of the implementation and the future direction of research.

Keywords:

• sampling
• stratified sampling
• probe vehicles
• traffic monitoring