论文部分内容阅读
为系统分析互通式立交分流区驾驶人的分流选择行为特征,讨论驾驶人多项分流选择行为的随机复杂性,根据车辆换道方式及位置对驾驶人分流换道行为进行了定量描述,确定了模型选择树及线性效用函数。选取分流区车辆行驶速度、加速度、车型、匝道小时交通量、减速车道车头时距等特征变量,建立了驾驶人匝道分流与分流方式选择的两层Nested Logit概率模型。基于广河高速公路部分互通式立交分流区的雷达跟踪数据,采用分阶段估计法进行模型参数估计。根据t检验判断了特征变量的影响程度,对模型进行了优化。结果表明:互通式立交区分流选择行为受多种因素综合作用的影响;两层NL概率模型具有较高的预测精度。
In order to systematically analyze the diversion behavior of drivers in the interchange overpass area, the random complexity of the driver diversion diversion behavior is discussed. The diversion behavior of drivers is quantitatively described according to the vehicle lane change pattern and location. Model Selection Tree and Linear Utility Function. A two-tier Nested Logit probabilistic model of diversion and diversion of drivers selection is established by taking characteristic variables such as vehicle speed, acceleration, hourly traffic volume of ramp, and headway of deceleration lane. Based on the radar tracking data of some interchanges in Guanghe expressway, the staged estimation method is used to estimate the model parameters. According to the t-test, the influence degree of the characteristic variables was judged and the model was optimized. The results show that the split flow selection behavior in the interchange is affected by the combination of many factors. The two-level NL probability model has higher prediction accuracy.