URFD: Urban Real-time Fatigue Driving Detection

Fatigue driving is a serious problem that probably causes a great damage to us. The fatigue driving action on the express way can be simply ruled according to the driving time, but it is ambiguous in the urban area and very hard to be detected. In this paper, we propose a reasonable definition of the fatigue driving action in urban area considering the driving time, driving speed, and driving area. According to the definition, we also design a distributed processing paradigm that can handle a massive volume of vehicle trajectories in real-time to detect the fatigue driving action. Finally, we conduct extensive experiments to evaluate the performance of our proposal based on a real data set released from DiDi.