基于车辆轨迹数据的城市道路驾驶行为分析

康佳宜，咸化彩^*，曾辉莉

山东交通学院交通与物流工程学院，山东 济南  250357

摘要：为提高城市道路交通安全，以车辆轨迹数据为基础，以济南市经十路某路段为研究对象，将道路划分为120个有效路段，研究道路坡度、交叉口、车道、公交车站、人行横道、辅道、交通吸引点等交通条件对车辆加速占比和减速占比的影响，选取主要影响因素建立驾驶行为与交通条件的多元Logistic回归分析模型，并选取长5 km的城市主干路为案例验证模型的适用性。结果表明：平坦路段、有交叉口路段更易产生加速和减速行为；有人行横道路段的车辆加速占比较无人行横道路段大，减速占比较无人行横道路段小；加速占比拟合回归模型和减速占比拟合回归模型预测结果准确率分别为73.91%、78.26%，模型适用性良好。

关键词：道路交通；驾驶行为；车辆轨迹数据；多元Logistic回归

Analysis of urban road driving behavior based on vehicle trajectory data

KANG Jiayi, XIAN Huacai^*, ZENG Huili

School of Transportation and Logistics Engineering, Shandong Jiaotong University, Jinan 250357, China

Abstract: In order to improve urban road traffic safety, based on vehicle trajectory data, taking a certain section of Jingshi Road in Jinan City as the research object, the road is divided into 120 effective sections, and the influence of traffic conditions such as road slope, intersection, lane, bus stop, pedestrian crossing, auxiliary road, and traffic attraction point on vehicle acceleration and deceleration proportion is studied. The main influencing factors are selected to establish a multivariate logistic regression analysis model of driving behavior and traffic conditions, and a 5 km urban main road is selected as a case to verify the applicability of the model. The results show that flat sections and sections with intersections are more likely to produce acceleration and deceleration behaviors; the acceleration proportion of vehicles on sections with pedestrian crossings is greater than that on sections without pedestrian crossings; the deceleration proportion is smaller than that on sections without pedestrian crossings. The accuracy of the prediction results of the accelerated proportion fitting regression model and the decelerated proportion fitting regression model are 73.91% and 78.26%, respectively, and the model has good applicability.

Keywords: road traffic; driving behavior; vehicle trajectory data; multivariate logistic regression