基于个体行为画像的路段拥堵责任划分方法

在学校、医院、商场等公共生活服务场所周边路段上经常发生的交通拥堵具有特殊性:是由特定事件触发特定出行者短时间扎堆聚集所导致的一种特殊类型的交通拥堵,这类拥堵的成因及治理对策具有重要的研究价值.利用自动车辆识别(automatic vehicle identification,AVI)技术对出行者的长、短期交通行为进行重构,探究导致公共生活服务场所周边路段的拥堵原因及拥堵责任划分方法.基于AVI数据建立了个体长、短期交通行为画像方法,通过多层卷积神经网络和层级聚类模型精确定位出导致公共生活服务场所周围路段拥堵的主要责任车辆.针对同一路段上的不同车辆,依据其长、短期出行行为特征,进行个性化精准管理,是在有限资源限制条件下解决公共生活服务场所周边拥堵问题的有效途径.选取安徽宣城市第六中学门前的常发路段拥堵作为研究对象.结果 表明:对于经过案例学校门前路段的所有出行者来说,只需对其中0.5％～0.7％的致堵车辆采取重点管理措施就可以有效缓解学校门前路段的拥堵问题.

Traffic congestion responsibility division method based on individual behavior portrait

It is a special type of congestion that periodically occurrs on the road around service centers such as school, hospital and mall. This kind of congestion is generally caused by periodic impulsive aggregation of specific travelers for certain events. The cause and relieving strategies for such congestion have both theoretical research and practical applicative values. In this study, the individual long-short term traffic behaviors were reconstructed based on automatic vehicle identification (AVI) technologies. The congestion around the service centers was identified through the reconstruction of the individual long-short term traffic behaviors. The vehicles that primarily responsible for impulsive aggregation congestion were precisely targeted via a proposed individual long-short term traffic behavior portrait framework, convolutional neural networks and prototype-based clustering method. According to the characteristics of long-short term travel behavior of different vehicles on the same road section, personalized precise management is an effective way to relieve the congestion around service centers under the condition of limited resources. The road management objectives were updated in AVI data environment and found that only 0.5% ~ 0.7% of the total number of vehicles passing by the case school require high-level management. It is an effective way to solve the impulsive aggregation congestion by formulating managements with different action levels and resolutions for specific travelers.