从驾驶员注视点变化的角度,应用SMI眼动仪,开展城市道路实车实验.采集驾驶员注意力特性,剖析在城市道路实车测试过程中驾驶员的注视行为,提出了基于马尔可夫过程的驾驶员视点转移模型.根据模型预测出驾驶员稳态下的注意力状态向量;并对驾驶行为进行预判.实车实验发现,驾驶员的视觉注意力在任一子域中停留时长服从对数正态分布;而视觉注意力离开某一子域的时长服从指数分布.机动车行驶过程中驾驶员主要注视正前方,其注视频率占49%,驾驶员关注右前方区域频次是左前方的1.6倍,对于正后视镜的关注多于左右后视镜.通过实车实验和视觉建模分析,能够更好地发现驾驶行为规律和不良的驾驶习惯,为驾驶环境和交通设施改善,保障交通安全提供支撑.%From a driver's visual point of view, real vehicle test was carried out using the SMI eye tracker in Beijing. Collecting and analyzing the characteristics of concentration and gazing behavior,a viewpoint transfer esti-mating model is proposed. Through the model,the state vector of attention in normal was calculated. Experiments found that the driver's visual attention staying at any sub domains obeyed the lognormal distribution,and visual at-tention away from a subdomain obeyed index distribution. The driver mainly gazed the directly in front of area in the driving process,and the frequency of fixation occupied 49%. The frequency of fixating in right viewpoint was 1.6 times more than left. There was more driver'attention in middle rearview mirror than left side mirror and right one. Through the real vehicle experiment and visual modeling analysis,the driving behavior rule and bad driving habits were found. And it is helpful to improve driving environment and traffic facilities,and provide the support to protect the safety of transportation.
展开▼
