考虑等待忍耐时间的电动自行车违规行为研判

Identification of E-bike violation behaviors by considering waiting tolerance time

引入生存分析法,将不违规和违规电动自行车骑行者的等待忍耐时间分别作为删失和完全数据,建立了电动自行车违规率函数,估算了单因素影响下的电动自行车违规率。将部分等待忍耐时间做分布形式假设,构建融合了参数方法和非参数方法特征的Cox比例风险回归模型,刻画多因素影响下的电动自行车违规行为。提取了1 335辆非机动车的等待行为数据,结合乘积极限法得到性别、年龄和非机动车类型影响下的电动车违规率函数。研究结果表明:中老年电动自行车骑行者违规率始终高于青年;当等待时间少于25 s时,女性电动自行车骑行者的违规率更高;当等待时间少于44 s或者在88~100 s时,电动自行车的违规率高于传统自行车和电动三轮车。与传统自行车违规行为的“从众”效应不同,当非机动车道狭窄或者是组群规模足够大,能够抑制电动自行车违规行为。增设协管员和专用左转相位能有效减少电动自行车违规行为。

By introducing the survival analysis,the waiting tolerance time of non-violation and violation electric bicycle riders was taken as deleted and complete data respectively.The violation rate function of electric bicycles was developed to estimate the violation rate under the influence of single factor.Secondly,part of the waiting time was assumed with a distributed form,and a Cox proportional risk regression model was proposed that integrated the characteristics of parametric and non-parametric methods to describe the violations of electric bicycle under the influence of multiple factors.The waiting behavior data of 1 335 nonmotor vehicles were extracted,and by using the product limit estimation method,the violation rate function of electric vehicles under the influence of gender,age and non-motor vehicle type was obtained.The results showed that the violation rate of middle-aged and elderly riders is higher than that of young riders;the violation rate of female riders is higher with the waiting time is less than 25 s;the violation rate of electric bicycles is higher than that of traditional bikes and e-tricycles when the waiting time is less than 44 s or between 88 s and 100 s.Unlike the "bandwagon" effect of traditional bicycle violations,when the non-motorized lanes are narrow or the group size is large enough,electric bicycle violations can be suppressed.The escort and dedicated left turn phase can effectively reduce the violation rate of electric bicycles.