Hamid Reza Eftekhari - Department of Computer, Faculty of Engineering, University of Malayer, Hamedan, Iran
Melika Zeraati - Student at University of Malayer

Urban traffic monitoring for predicting congestion and managing traffic is one of the concerns of municipal managers. In this study, using user’s smartphone’s Wi-Fi, are proposed a cost efficient approach to collect traffic data and to detect the traffic condition of the roads. Firstly, the number of users is collected by a Wi-Fi scanner during each one-hour intervals. After extracting the feature from thesedata, the traffic conditions are estimated by different supervised learning machine including Artificial Neural Network (ANN), Decision Tree (DT), Support Vector Machine (SVM), K-Nearest Neighbors (KNN) and Naïve Bayes (NB). Finally, the threshold is calculated using thebest cut-off point in Receiver Operating Characteristic (ROC) curve. The results show that KNN algorithm with the accuracy of 93.3% is the best estimator. The precision of traffic congestion detection in this algorithm is 92.1%. Also, the threshold was 38 for the case study. This threshold can be shown beginning of congestion in the road and can be used for applying traffic policies to management conditions

Road Traffic monitoring; Intelligent Transportation System; Wi-Fi Scanner; Smartphone