Internet-of-vehicle-oriented fatigue driving state detection mechanism

Fatigue driving is one of the most common dangerous driving state. When a driver is fatigue, his or her reaction speed will be reduced, which will easily lead to traffic accidents if not manipulated in time. It is an effective way to realize the detection and adjustment of the driver's fatigue state based on the Internet of vehicle. Current fatigue driving state detection mechanisms usually have some obvious defects, including low detection accuracy rate, high hardware cost, uncomfortable wearing experience, high environmental requirement, etc. We propose a novel Internet-of-vehicle-oriented fatigue driving state detecting mechanism based on data fusion with the Dempster-Shafer (D-S) evidence theory. We apply the D-S evidence theory to combine the above two initial detection results to realize the decision level fusion, in order to get a better result. We design and construct the prototype system of the Internet-of-vehicles-oriented fatigue driving state detection. The experimental results show that, compared with the existing methods, the proposed mechanism has the certain advantage in the detection accuracy rate. We also implement a series of experiments to compare three fatigue driving state detecting mechanisms from the response time, time complexity and memory consumption.