| In recent years,major and serious traffic accidents happen frequency, driver fatigue is one of the important reasons. Statistics of 2009 from Ministry of Public Security show that about 15.2% of the traffic accident deaths due to the fatigue driving, it can be concluded that it is the great harm of fatigue driving. Relative research indicates that, compared with non-fatigue condition, It shows a significant difference of driver in the aspects of physiological and external behavior under the fatigue status. Therefore, it is of great theoretical significance and practical value of doing research about drivers'physiological and eye movement characteristics under the fatigue state in analyzing the degree of driver fatigue, and improving the safety of road traffic.Driver fatigue is defined as the phenomenon of the fall of physical function caused by the cumulative and superposition of driver fatigue based on the original fatigue. According to experimental conditions and effectiveness and reliability of test indicators on the reflection of driver fatigue, three typical brain waves and heart rate variability indexes are chosen as the indexes for the physiological measurement, and blink duration, blink frequency, the percentage of fixation time, the number of fixation point, saccade amplitude and the average speed of saccade are chosen as the indexes of eye movement. Experimental program is designed and carried out from two aspects of static and dynamic.In the light of data of the experimental test, changing rules of variation coefficient of the three typical brain waves are analyzed. Variation regulations of the linear and nonlinear characteristics of heart rate variability (HRV) indexes are analysed. Quantitive relationship among the amount of original fatigue, cumulative amount of driving fatigue and fatigue accumulation is analyzed based on the variation coefficient of R-R interphase. The S-curve model of driving fatigue accumulation is established for describing the diversification of driving fatigue accumulation over the time. Physiological fatigue degree is defined for determining the key indicator of physiological characteristic measurenment on the status judgement of driver fatigue.In line with data of the experimental test, the measure indicators'variation regulation of the three eye movement types which are blink, fixation and saccade are analyzed. Transfer rule of the driver'fixation point is analyzed by calculating the one step transition probability matrix of markov chain. The gaze distribution coefficient of AOI (area of interestion) on the driver vision field is put forward, and the variation coefficient of fixation distribution is calculated for analyzing the quantitive relationship among the amount of original fatigue, cumulative amount of driving fatigue and fatigue accumulation. The S-curve model of driving fatigue accumulation is established for describing the diversification of driving fatigue accumulation over the time. Eye movement fatigue degree is defined for determining the key indicator of eye movement characteristic measurenment on the status judgement of driver fatigue.Comparing the advantages and disadvantages on the various methods of information fusion, combination with the the input-output process of driver fatigue, BP neural network is chosen as the approach of multi-feature information fusion for judgement driver fatigue status. BP neural network models of driver fatigue are constructed separately on the condition that driving physiological fatigue degree and driving eye movement fatigue degree are considered as a single input respectively and both of them are taken as the simultaneous input. The model is verificated with data measured by experiment. The result shows that BP neural network model which driving physiological fatigue degree and driving eye movement fatigue degree are simultaneous inputted can judge accurately non-fatigue, mild fatigue and severe fatigue. The model can be applied to fatigue driving monitoring. |
PDF Full Text Request