基于道路交通环境的非机动车骑行行为特征研究

Study on Non-motorized Vehicle Riding Behavior Characteristics Based on Road and Traffic Environment

为促进路段非机动车服务水平评价及骑行环境改善,以道路交通环境为基础,研究非机动车骑行行为特征是十分必要的.以机非实体隔离、未施划人行横道的非机动车专用道为研究对象,在骑行行为特征影响因素分析的基础上,针对具有不同道路交通环境的典型路段采集非机动车骑行行为信息,应用二维运动图像解析技术与数理统计分析方法研究速度、加减速度、加速度变化率、横向偏移率等骑行行为特征,剖析不同道路交通环境下骑行行为特征的特点;在此基础上,利用广义回归神经网络建立基于道路交通环境的非机动车骑行行为特征模型.通过模型精度分析,证明基于道路交通环境的广义回归神经网络模型对非机动车骑行行为特征的预测效果良好.

In order to promote the level of service evaluation and riding environment improvement of non- motorized vehicles, it is necessary to study non-motorized vehicle riding behavior characteristics based on road and traffic environment conditions. This paper specifically studied non-motorized vehicle lanes physically separated from vehicular traffic and without crosswalk markings. Based on the analysis of riding behavior influence factors, the paper collected information of non-motorized vehicles riding behavior on typical sections with different road and traffic environment conditions. It researched on riding behavior characteristics such as speed, acceleration, acceleration rate of change, lateral movement using the 2D motion image analysis technology and mathematical statistics analysis method. It analyzed the riding behavior characteristics under different road and traffic environment conditions. Based on the road and traffic environment conditions, we finally established non-motorized vehicle riding behavior characteristic model by using the general regression neural network. The results show that general regression neural network （GRNN） model can better predictnon-motorized vehicle riding behavior characteristics based on road and traffic environment conditions.