Evaluation of accuracy of the infra-red traffic logger under mix traffic condition

Abstract

Traffic data collection under heterogeneous traffic conditions is one of the major challenges faced by highway agencies, policy makers as well as researchers. The Infrared Traffic Logger is a highly advanced, non-intrusive traffic surveillance system which can detect different classes of vehicles based on wheel base size and axle spacing. The fundamental macroscopic traffic parameters: flow, speed and composition under different traffic conditions and weather condition were checked against to base line condition. It was found that overall accuracy of vehicle counts from infrared traffic logger is more than 90% except for two-wheelers under heavy traffic condition and for heavy vehicles under medium and heavy traffic condition. Speed detection accuracy of infrared traffic logger also more than 90% and accuracy of vehicle classification also more than 95% except multi axle vehicles. Multi axle vehicles are misclassify as number of twoaxle vehicles under medium or high traffic conditions. Correction factors are suggested to adjust the accuracy level for over counting and under counting situations.