基于注视行为特性的驾驶人分心负荷评估

The Evaluation of Driver's Distraction Load Based on Fixation Behavior Characteristics

  • 摘要: 为了评估驾驶人操作手机的各种行为对行车安全的影响,组织19名驾驶人佩戴眼动仪开展室内模拟驾驶试验,分别采集城市快速路自由流和拥挤流2种典型交通状况下驾驶人进行正常驾驶、免提通话、语音短信3种操作时的注视数据。采用层次聚类法结合机械划分的方式,将驾驶人的视野平面划分为6个视觉兴趣区域;利用描述统计和方差分析法建立了衡量注视行为特性的敏感性指标集(注视区域信息熵、注视持续时间、垂直方向注视偏差、瞳孔面积变异系数);最后采用熵权法构建4项指标的权重体系,提出了分心负荷指数的概念,并引入TOPSIS法验证了分心负荷指数对分心程度的评估效果。结果表明:进行免提通话操作时驾驶人多处于认知分心状态,语音短信操作时多处于视觉分心状态;相较于正常驾驶,除了自由流场景中进行免提通话操作时分心负荷差异甚微以外,其余手机操作均对驾驶人注视行为产生显著影响,致使分心负荷指数普遍升高,且在拥挤流场景中,执行语音短信操作时分心负荷激增,驾驶风险远高于免提通话。

     

    Abstract: In order to evaluate the impact of various behaviors of drivers operating mobile phones on driving safety, 19 drivers were organized to wear an eye tracker to carry out indoor simulation driving test, and the fixation data of drivers during normal driving, hands-free call and voice message in two typical traffic conditions of free flow and crowded flow of urban expressway were collected respectively. The driver's visual field plane is divided and classified into six visual interest areas with the hierarchical clustering method combined with mechanical division. Using descriptive statistics and variance analysis, a set of sensitive indexes (information entropy of fixation area, fixation duration, vertical fixation deviation and coefficient of variation of pupil area) was established to measure the characteristics of fixation behavior. Finally, the entropy weight method was used to construct the weight system of four indicators and propose the concept of distraction load index, and the TOPSIS method was introduced to verify the evaluation effect of distraction load index on the degree of distraction. The results show that most drivers are in the cognitive distraction state during hands-free call operation and in the visual distraction state during voice message operation. Compared with normal driving, except that there is little impact on distraction load during hands-free call operation in the free flow scenario, other mobile phone operations have a significant impact on the drivers' fixation behavior, resulting in a general increase in distraction load index. In addition, in the crowded flow scenario, the distraction load increases sharply when performing voice message operation, and the driving risk is much higher than that of hands-free call.

     

/

返回文章
返回