中文摘要：

为提高车内次任务影响试验的标准化与绩效指标适用性,在应用换道测试（LCT）国际标准试验环境的基础上,增加驾驶绩效指标,使用基于C4.5决策树的数据挖掘方法研究多个指标的综合特性。试验中,选择配对记忆试验作为次任务,在基本任务、简单任务和复杂任务3种任务条件下,采集车辆行驶轨迹、车速、转向角等数据进行分析。结果表明：车道偏移次数、初始换道距离、正确换道比例等新增加的指标能够从横向偏离超阈值事件次数、驾驶人反应时间、失误比例等方面反映新的驾驶特性;应用C4.5决策树模型,分别使用车道保持指标组和换道控制指标组对驾驶状态进行区分,可以反映次任务对指标组整体的影响,而非对单个指标的影响;应用该方法对车道保持和换道控制驾驶状态的正确分类比例分别达到85.2%和91.3%,表明次任务对驾驶人车道保持综合水平与换道控制综合水平的影响非常明显。

英文摘要：

In order to improve the standardization of in-vehicle secondary task influencing experiment and applicability of performance indicator,driving performance indicators were added based on the utilization of lane change test（LCT）international standard experiment environment.Data mining method built on C4.5decision tree was applied to analyze comprehensive characteristics of multi indexes.Paired memory experiment was chosen as secondary task while vehicle traveling trajectory as well as speed and steering angle was collected in three conditions separately including baseline,simple task group and complex task group.The results show that the additional new indexes such as lane outlines and initial lane changing distance as well as correct proportion of lane changing can reflect new driving characteristics from three aspects,which lie in time for the lateral deviation exceeding the threshold event and reaction time of drivers as well as error ratio.Based on C4.5decision tree,the lane keeping indicator group and lane changing indicator group are used to discriminate driving state,which can reflect effect of secondary task on the whole indicator group,not olny the single indicator.The accurate classification proportions reach to85.2% and 91.3% respectively,which shows that secondary task has a significant influence on lane keeping and changing.