Research On Fatigue Driving Behavior Detection System Based On Acoustics

Dangerous driving due to drowsiness and distraction is the main cause of traffic accidents,causing heavy casualties and economic losses,and is becoming an increasingly serious problem worldwide.Therefore,it is urgent to solve this problem by accurately detecting dangerous driving behaviors and generating real-time warnings.Inspired by the observation that various dangerous driving behaviors cause the common acoustic characteristics to respond to the sound source signal,this paper proposes a driving sonar system.The system accurately detects dangerous driving behavior and uses off-the-shelf smartphones to generate real-time alerts.Compared with the most advanced systems,this article does not require special sensors,only the speakers and microphones built into the smartphone.Specifically,the system can accurately recognize head movement and steering patterns,such as head nodding,yawns,and sudden steering wheel adjustments,and can recognize different drivers' head movements and steering wheel steering patterns,with only a small amount of training.,The main work of this article is as follows:1.We have proposed and designed a dangerous driving recognition system that does not require the use of additional dedicated sensors,and ordinary commercial smart phones used in the market can meet the requirements of the system.As far as we know,we are the first system to use acoustic echo to detect dangerous driving behavior.2.Through observation,the essential characteristics of dangerous driving behavior can be expressed as the movement pattern of the head and steering hand.We use the motion characteristics of the driver's head and steering hand to determine the dangerous driving action together,and propose a two-state processing model to deal with fatigue driving separately And distracted driving.3.In order to test the system,we conducted extensive experiments in simulated and real driving environments to test the robustness and accuracy of the algorithm.The entire system is divided into two parts: a data collection module and a background algorithm execution module.Effectively verify the correctness of the algorithm.