Using ETC Data and Detector Data to Derive Information of Route Choices on Tokyo Metropolitan Expressway (MEX)

Route choice analysis based on revealed preference (RP) data plays an important role in various applications on urban expressway. Yet, the exact route choice information is difficult and expensive to be obtained in practice. Conversely, the availability of traffic surveillance systems and electronic toll collection (ETC) system opens possibility to derive route choice information for each vehicle. This paper focuses on the field evaluation of two route identification principles: non-parametric and parametric method, introduced by Pelata et al. (2008). The paper also modifies the parametric method using the more realistic assumption in probability distribution. The assessment of these methods is based on the dataset from four study cases on urban expressway in Tokyo. The paper finds that both methods provide little differences in performance. In most cases the non-parametric method is found to perform marginally better than the other. The accuracy of route choice information derived from both methods ranges between 60 and 75%