The development and implementation of driving behavior analysis system

The occurrence of traffic accidents usually results from drivers’ bad driving behavior. By analyzing drivers’ bad driving behavior, it can help to avoid dangerous driving conditions, and also reduce traffic accidents. A system has been designed in this system that could simultaneously collect images of the roads ahead, and real-time images of the head, eyes, and face of the driver while driving. There were three subsystems constructed in this system: (1) Driver's visual state analysis system, (2) ECU message capturing and decoding system, (3) Driving behavior analysis system.Meanwhile, through integrating the information of the driver’s visual state analysis system with ECU message capturing and decoding system, the current system could detect and identify four types of bad driving behavior: (a) turning without flashing the turn signals, (b) Not looking in the rearview mirror when turning, (c) Distracted driving and (d) Fatigue driving. On top of that, the accuracy rates of "turning without flashing the signals" and "without looking in the rearview mirror when turning" were 89% and 87%, respectively, while the accuracy rates of distracted driving and fatigue driving were 82% and 79%, respectively.