Vehicle Detector Evaluation Based on Traceability and Confidence Interval Concepts

摘要

Vehicle detectors are an essential device for advance traffic management and information system(ATMIS). Hence, the accuracy of traffic data from detectors is of fundamental importance. Whenevaluating detectors, one uses a baseline data source that generates reference data. However, referencedata also contain errors and uncertainties, albeit small ones. Nevertheless, little research has beenperformed to address the errors and uncertainties of reference data. This study proposes a new detectorevaluation scheme to resolve the problem. The scheme takes errors and uncertainties of reference datainto account for reliable detector evaluation. Moreover, it applies the interval estimation technique,instead of the traditional point estimation, to present detector evaluation results. The suggested scheme isthoroughly verified using the Korea detector test bed, and resultantly produces reliable detector evaluationresults.