基于马尔可夫过程的路网交通状态预测方法

摘要

针对城市道路交通管控对路网实时交通状态信息的需求,首先对交叉口的交通状态进行了界定并划分为4个状态等级,然后以15 min为时间间隔,结合历史数据采用马尔可夫过程计算每个交叉口的状态转移概率,并对各个交叉口下一时刻的交通状态进行预测,进而对每个交叉口进行分类并采用综合评价法确定整个路网的交通状态.试验以MAPE为评价指标对已有算法进行对比,结果表明:该方法不仅能够实时地确定每个交叉口的交通状态,同时能够基于不同类型交叉口的预测结果确定整个路网的交通状态,且交叉口平均预测误差为7.04%.%Aiming at demands of urban traffic control and management for real-time traffic status of road network, this paper firstly gives the definition of traffic state of an intersection and divides the states into 4 levels.Then the state transition probability of each intersection is computed by using Markov process combined with historic data at time intervals of 15 mins, and the next state of each intersection is predicted.Finally, each intersection is classified and a comprehensive evaluation method is used to predict the whole traffic states of a road network.Experimental results show that this method can not only compute the real-time traffic states of each intersection, but also can give the traffic states of the whole road network based on prediction results of different types of intersections, and the average forecast error at the intersection is 7.04%.