Fatigue driving identification method based on driving behaviors and eye movement characteristics

Abstract

The invention discloses a fatigue driving identification method based on driving behaviors and eye movement characteristics. The method comprises the steps that 1) a fatigue driving simulation experiment is designed, and the driving behavior data of a driver and the eye movement data of the driver under different fatigue states are collected through the experiment; 2) the experiment data are segmented and screened to establish a fatigue driving sample database; 3) a single-factor variance analysis method is used to carry out statistic analysis on the significance of the difference of driving behavior parameters and eye movement characteristic parameters of the driver under different fatigue states, and the optimal time window of each characteristic parameter is found out; 4) bivariate Spearman correlation analysis is carried out on initially selected characteristic parameters; 5) the optimal characteristic parameters are selected; 6) a BP neural network identification model for fatigue driving is established; and 7) Matlab software is used to write a model program, and a training sample and test sample set are randomly selected to train the model and identify a fatigue state.

Description

technical field [0001] The invention relates to a fatigue driving identification method based on driving behavior and eye movement characteristics. Background technique [0002] With the rapid development of the transportation industry, the number of motor vehicles is constantly increasing, and the number of road traffic accidents is increasing. The problem of road traffic safety has become a serious social problem. Fatigue driving is one of the important causes of traffic accidents. In recent years, there have been more and more research topics on fatigue driving, and fatigue driving detection technology has also been developed rapidly. Existing research has achieved good results in fatigue detection. But there are still deficiencies: [0003] 1) It is a gradual process for the driver to change from sobriety to fatigue. The current research often ignores this process. It is impossible to judge when the driver enters the fatigue state, and it is impossible to give reasonabl...

AI Technical Summary

Problems solved by technology

[0003] 1) It is a gradual process for the driver to change from sobriety to fatigue. The current research often ignores this process. It is impossible to judge when the driver enters the fatigue state, and it is impossible to give reasonable prompt information for the actual situation of the driver. Therefore, reasonable It will be a focus of research to accurately classify and accurately identify the driver's fatigue state;

[0004] 2) Existing detection methods based on a single indicator cannot overcome the influence of space, light, weather and other environments on detection accuracy. Therefore, the use of detection methods based on information fusion will be an important way to improve detection accuracy and reliability

Images