Abstract:

Human factors are among the causes of traffic accidents. Especially in large cities, the number of drivers and vehicles increases the risk rate in traffic. In everyday life, driver behaviors vary according to controlled driving. Therefore, it has become an important topic to identify the behavior of drivers during driving in the prevention of traffic accidents. Driving data can be collected with the help of devices and sensors, such as the smartphone, camera, drive meter, gyroscope, control field network, radius detection distance measurement, inercial measurement unit, to detect drivers’ behavior. To identify and evaluate driving, it is necessary to take into account different biological and physiological factors such as driver psychology, environmental factors, road condition, vehicle kinematics. This study used driving data from two different data sets recorded in real traffic environments. On both data sets, the U.S. National Road Traffic Safety Administration has identified severe acceleration and maneuver events and conduct identification has been performed. Developed web-based software and the features extracted based on the data obtained from the tool are presented graphically and statistically. Furthermore, travel-related routes, speed, stop and stop events are visualized to identify driver behavior with user-friendly interface on the map. At the end of the drive, the driver's score graph and the results of the drive report are created. Driving can be easily assessed. The work is a low-cost driver analysis application because it is done with open source software.

Keywords: