城市快速路实时交通状态估计和行程时间预测

根据城市快速路交通诱导和监控系统的实际需要,提出了实时估计和预测城市快速路上交通状态和任意两点间动态行程时间的方法.其基本思想是将扩展卡尔曼滤波理论引入宏观动态交通流模型,结合快速路上的固定检测设备,实时估计和预测未来几个时段的交通状态,并利用“虚拟车”法预测动态的行程时间.通过对上海市快速路典型实测数据的实例分析,发现交通状态估计模型具有良好的跟踪能力,行程时间预测模型在畅通状态计算结果和实测结果几乎完全重合,拥挤状态相对误差基本维持在10%以下.结果表明,该模型的适用性和精度都令人满意,可为城市快速路交通控制和诱导提供依据.

Real-Time Traffic State Estimation and Travel Time Prediction on Urban Expressway

According to the need of traffic guidance and traffic monitoring system on urban expressway,the paper introduces a method to estimate and predict traffic states and travel time between arbitrary locations on urban expressway.The basic concept is to estimate future traffic states using macroscopic dynamic traffic flow model integrated with extended Kalman filtering(EKF)and the fixed traffic detectors on urban expressway,and to predict dynamic travel time using "fictitious car" method.The method is tested on urban freeway in Shanghai by using real-detected data,in which the traffic state estimation model shows good tracking ability,and results of travel time prediction model show that there is a slight difference between the calculated result and the real in free flow,and a relative error at about 10% mostly in congested flow.The precision and applicability of this mothod are acceptable,and it can be used to provide a basis for traffic control and traffic guidance.