Traffic Anomaly Detection on Chalerm Mahanakhon Expressway Using Web-Based Traffic State Data

Abstract

Detection of traffic anomalies is useful in areas where traffic congestion is an issue of concern. This study aims at developing a system to detect traffic anomalies occurring on the Chalerm Mahanakhon expressway in Bangkok by using web-based traffic state data. In this study, artificial traffic sensors were placed along the expressway to analyze the traffic state data. A statistical model was developed to predict the traffic states on these artificial sensors based on Bayes’s theorem. An algorithm to detect traffic anomalies was introduced. The degree of anomaly was defined by comparing the probability of the expected traffic state and the probability of the observed traffic state. The method can be applied to any area where sufficient historical data are available.