【作者】 张琦；

【导师】 吴超仲；

【作者基本信息】 武汉理工大学 ， 交通运输工程， 2018， 硕士

【摘要】 疲劳驾驶行是道路交通中的常见现象。驾驶过程中,驾驶人会逐渐积累自身疲劳,当其疲劳程度达到一定范围后,驾驶人对外界信息的提取并处理的速度、自身的反应速度都会有所降低,甚至可能在驾驶过程中陷入沉睡,这都会在很大程度上导致严重交通事故的发生。随着道路交通逐渐提速,在单一道路环境下连续驾驶,驾驶人疲劳累积速度也会增加。同时驾驶期间休息时长、驾驶前睡眠时长以及驾驶期间的昼夜节律变化情况都会实时影响到驾驶人的实际疲劳程度,进而导致驾驶疲劳事故的发生。在综述了驾驶全周期疲劳程度时变规律的相关理论及国内外研究现状后,从实验数据出发,结合现有疲劳累计框架模型拟合出相应因素的疲劳预测模型,并将模型应用到智能手机中。本文的具体研究内容如下:首先,根据驾驶疲劳组成成分,针对影响驾驶人疲劳的各项因素,选择相应的疲劳评定指标、行驶状态检测设备等,设计相应的实车实验,最终招募了34位被试驾驶人,其中24名被试进行上午组实验,各有5名被试进行下午组及夜间组实验,同时从3组实验中各随机选择一名驾驶人作为验证者。其次,在收集到驾驶人驾驶全周期的疲劳变化数据之后,针对初始数据进行相应的插值、拟合处理。在昼夜节律对应的疲劳数值变化确定后,对驾驶总疲劳进行依次削减的方法,依次获得睡眠、驾驶、休息时长三因素对应的疲劳值变化情况。之后,对31位被试数据的四个因素对应的时间-疲劳值变化模型进行拟合,获取最终驾驶全周期疲劳程度时变规律预测模型,并使用SDLP这一指标对验证者相关数据进行验证。结果展现本文模型的准确率较高。最后,文章将驾驶全周期疲劳程度时变规律预测模型应用到智能手机端,利用智能手机对驾驶过程中驾驶人的疲劳程度进行预测,并在驾驶人疲劳超过阈值时,对其进行提示。通过对驾驶全周期疲劳程度时变规律预测模型的建立,可以对驾驶人实时疲劳程度进行预测并进行实时提醒,以期减少疲劳事故的发生,提高道路交通环境安全性;并为疲劳驾驶相关法规的优化提供理论支撑。更多还原

【Abstract】 Fatigue driving behavior is a kind of driving behavior that is very easy to see in the road traffic.In the driving process,the fatigue will accumulate gradually,when the fatigue degree reaches a high level,the speed of the driver’s extraction,processing and response of information from the outside will be reduced,even fall asleep in the driving process,all will lead to serious traffic accidents to a great extent.With the speed of vehicle in road transport increase gradually,the cumulative speed of driving fatigue will increase while driving under single road environment.Meanwhile,during the driving period,the rest time,the quality of sleep before driving and the circadian rhythm during driving will affect the actual fatigue degree of drivers in real time,which will lead to driving fatigue accidents.After summarizing the related theories and the research situation of the whole cycle fatigue time-varying rule at home and abroad,the corresponding fatigue prediction model was fitted out based on the experimental data,and then the model was applied to the smart phone.The specific researches are as follows:First of all,according to the components of driving fatigue,in view of the various factors that influencing driving fatigue,the corresponding fatigue evaluation index,and driving condition detection equipment were selected,the field experiments were designed.Finally 34 pilot drivers were recruited including 24 subjects in morning experimental group,5 subjects in afternoon group and 5 in night group,and three verifier from each group were randomly selected.Secondly,after the driver’s full cycle fatigue changing data are collected,the corresponding interpolation and fitting were carried out for the initial data.After determining the corresponding fatigue changes in the circadian rhythm,the method of decreasing the total driving fatigue in turn was used to get the change of the fatigue value corresponding to the duration of sleep,continuous driving and rest time.After that,four factors corresponding to the 31 subjects’ data were fitted to the time-fatigue value changing model,and finally the prediction model of the time-varying rule of the full driving cycle fatigue level was obtained.And the SDLP was used to verify the relevant data of the verifier,the results show that the prediction model has a high accuracy.Finally,the fatigue prediction model was applied to the smart phone,and the prediction of fatigue degree change during the driving process is completed by smart phone,meanwhile,the driver is prompted when the driver fatigue exceeds the threshold.Through the establishment of variation prediction model on driving fatigue level,the driver fatigue could be predicted and the driver could be alerted,reducing the incidence of fatigue accidents,improve road traffic safety,and provide theoretical support for the optimization of relevant laws and regulations.更多还原