Formulation of a new methodology to identify erroneous paired loop detectors

Abstract

A new algorithm was developed for screening data from paired inductive traffic loop detectors. The algorithm estimates the effective vehicle length for each vehicle that passes the paired loop detectors and computes the average vehicle length for a fixed set of vehicles, the size of the set depending upon the vehicle volume flow. Error detection is done in two steps: 1) to calibrate the effective average vehicle length range for a detector; and 2) to compare this to the effective average vehicle length computation online.