Automatic Traffic Accident Detection Based on the Internet of Things and Support Vector Machine

Abstract

Recent years, the transportation efficiency and related issues have become one of the main focuses of the global world. Along this line, intelligent transportation systems (ITS) based on Internet of Things (IoT) has become a common solution to may transportation problems, such as traffic prediction, road status evaluation, traffic accident detection, etc. In this paper, we focus on the traffic accident detection in the IoT platform. Specifically, we propose an overall framework of intelligent transportation. Then, for traffic accident detection, we proposed a Support Vector Machine (SVM) modified by Ant Colony Algorithm (ACA) as the solution. We conduct experiments on real world traffic data to predict 7 types of traffic accidents. The results show that our proposed method is effective and efficient.