Lifespan Evaluation of Traffic Detector for Automated Traffic Recorders Based on Weibull Distribution

Abstract

An automatic traffic recorder (ATR) makes it easy to detect a vehicle’s speed and the number of axles because it produces highly precise data; thus, it can be used to obtain major traffic characteristics for several purposes such as vehicle classification. However, the failure of a piezo sensor or inductive-loop sensor often results in missing data. Moreover, there is no method available for evaluating the life of a sensor installed in a pavement. Therefore, such sensors require maintenance after a failure occurs, resulting in a large amount of missing traffic volume data. If the quantified life expectancy of a sensor could be obtained, a maintenance and rehabilitation strategy could be established and the economic feasibility could be determined. In this study, the 5-year maintenance history of a sensor newly installed in a national highway was reviewed. Normal, gamma, lognormal, and Weibull distributions were fitted to the data, and the results were compared. The Weibull distribution was used to predict the life. When the maintenance histories of piezo and inductive-loop sensors were fitted to the Weibull distribution, the scale parameters were 79.23 and 89.72 and the shape parameters were 3.28 and 2.83 for the piezo sensor and loop sensor, respectively. The mean life and median life of the piezo sensor were calculated to be 71.05 and 70.85 months, respectively. Those of the loop sensor were 79.92 and 78.82 months, respectively.