An intelligent personalized traffic information extraction system for road traffic safety

Other than some driving assistant systems that can automatically avoid accidents, providing a driver with highly relevant and real-time traffic information is useful in attracting a driver's attention and striving more reaction time to possible dangers. In this paper, we propose an intelligent traffic information extraction system that explores a vehicle's trajectories to discover its driver's movement patterns and use the discovered patterns to predict the most likely locations that the driver will go in the near future. Based on the proper locations in the near future, our system extract the top-k correlated traffic messages that are situated on the proper way of the driver. To validate our design, we implement the intelligent traffic information extraction system as an Android app and run the app on a car to test the system. The results show the discovered movement patterns can help in extracting highly correlated traffic messages and as the movement routes of a driver are of high regularity, more percentage of the extracted traffic events are situated on the way of the driver.