基于数据融合的城市道路交通事件自动检测

Automatic Traffic Incident Detection for Urban Roads Based on Data Fusion

针对交通事件自动检测多以高速公路、城市快速路为对象以及使用数据源单一的现状,提出一种基于多源数据融合的城市道路交通事件检测方法。在对信号控制下交通事件引起的交通流变化进行分析的基础上,利用杭州市城区浮动车、SCATS、Citilog系统提供的实时交通数据,基于CUSUM算法构建差分流量和速度交通事件检测模型。该模型可以有效抑制交通信号对于交通流的周期性影响。实验表明,模型在高峰时段和平峰时段均能快速准确检测交通事件。

Traditional automatic traffic incident detections（ATID） mainly focus on freeway and expressway,and use single data resource,such as loop data.In this regard,this paper proposes an approach for ATIS on urban roads based on multiple data fusion technology.This paper analyzes the change of traffic flow caused by incidents under traffic signal control.On the basis of the data from floating car,SCATS,and Citilog systems in Hangzhou City,an ATID model uses the CUSUM algorithm to calculate differential flow rates and speeds.The model can effectively suppress the periodic impact from traffic signals on traffic flows.The experimental results show that,during both peak hours and non-peak hours,the proposed model can detect traffic incidents more rapidly and accurately.