Detection Method of Driver's Non-Normal Physiological States for Automobile Safety
Currently autonomous driving has been developed in many countries, which may enhance safety performance of automobile in combination with artificial intelligence. Driver’s non-normal physiological state has higher potential risk which may be involved in traffic accidents. Driver’s physiological states adaptive driving safety function is thought one of promising candidates to be introduced into autonomous driving system. This research reviewed driver’s non-normal physiological states by analyzing real world traffic incidents data collected by Internet survey. Then this research focused driver’s distraction and drowsiness as well as anger which may have potential risks to be involved in traffic accidents. Heart beat and eye movement are often identified as alternative characteristics of driver’s distraction. Also, facial expression is identified as alternative characteristics of drowsiness and anger. Signal processing may be indispensable technology to detect driver’s non-normal physiological states. This research applied image processing to detect driver’s distraction and facial expression processing to detect driver’s drowsiness and anger. Finally, this research proposed an example of driver’s physiological states adaptive driving safety function in combination with artificial intelligence which may play important role in autonomous driving system.
Keywords - Driver’s Distraction, Drowsiness, Anger, Autonomous Driving, Artificial Intelligence, AdaBoost, Neural Network.