摘要
为更深入研究无信号控制路段行人过街的交通特性,通过激光雷达采集行人过街过程中来车的运动状态数据,考虑外部因素对行人过街决策的影响及行人对安全过街的心理需求,利用人车间距、车辆速度和可穿越间隙等参数分析直接过街和等待过街2种决策结果下的数据特性。基于上述参数建立行人过街安全心理距离模型和多元Logistic回归决策模型,并验证模型有效性。结果表明:当人车间距越大、车速越低、可穿越间隙越大时,行人过街概率越大;相同可穿越间隙下,车速越快,行人过街概率越大;所建立二模型均有较高有效性。研究成果可为车载行人预警系统的优化提供依据。
In order to study the traffic characteristics of pedestrian crossing the street without signal control further,data on motion of incoming vehicle during the process of pedestrian crossing the street were collected by using a Lidar,and effects of external factors on pedestrian crossing decision and those of the psychological needs of pedestrians on safe crossing were taken into consideration.The parameters,including the distance between the pedestrian and the incoming vehicle,the velocity of the vehicle and the gap for crossing,were analyzed for two decision results,crossing directly and crossing after waiting.The above parameters were used to build a psychological distance model for the pedestrian crossing the street safely and a multivariate Logistic regression decision model.The effectiveness of the two models was evaluated.The results show that the probability of crossing the street for the pedestrian increases when the distance between the pedestrian and the incoming vehicle grows,the velocity of the vehicle decreases,and the gap of crossing becomes larger,that with the same gap of crossing,as the velocity grows,the probability of crossing gets larger,that the effectiveness of the models is great,and that the research results can provide a basis for the optimization of the on-board pedestrian warning system.
作者
赵佳
宋柱
张名芳
王畅
ZHAO Jia SONG Zhu ZHANG Mingfang WANG Chang(School of Automobile, Chang＇an University, Xi＇an Shaanxi 710064, China)
出处
《中国安全科学学报》
CAS
CSCD
北大核心
2017年第5期25-30,共6页
China Safety Science Journal
基金
"十二五"国家科技支撑计划项目(2014BAG01B05)
陕西省自然科学基金资助(2016JQ5096)
中央高校基本科研业务费专项资金项目(310822161009
310822171118
310822172001)
关键词
过街决策
无信号路段
行为特性
可穿越间隙
多元LOGISTIC回归
crossing decision
without signal control
behavioral characteristics
acceptance gap
multivariate Logistic regression
