论文部分内容阅读
为预测车辆在交叉口群的旅行时间,首先利用城市交通系统呈现的分布、并发特性,采用时延赋色Petri网(TCPN)进行模块化、层次化的建模,建立了包括输入/输出路段车流、交叉口车流、信号控制,以及交叉口群车流的TCPN模型;其次,利用模型的监视器获取仿真过程中车辆的紧迫程度、进出交叉口群的时戳、进出指定库所的时戳、库所容量等状态信息,在此基础上提出了基于紧迫程度和车流密度的车辆平均速率模糊推理算法,实现了对自由旅行路段上车辆平均速率以及车辆旅行时间的预测。试验结果表明:采用基于TCPN交通流模型和模糊推理相结合的仿真预测方法能够合理地对交叉口群中车辆旅行时间进行预测,并且相对于卡尔曼滤波预测方法,提出的方法平均绝对百分误差和均方根误差累计值分别降低了14.33%和22.98%。
In order to predict the travel time of vehicles at intersections, the paper first makes use of the distribution and concurrency features presented by urban transport system, and uses the time-varying colored Petri nets (TCPN) to modularize and hierarchize the model. Traffic flow, traffic flow at intersection, signal control and TCPN model of traffic flow at intersection; secondly, using model monitor to obtain the urgency of the vehicle in the process of simulation, timestamp in and out of intersection group, timestamp in and out of designated place, Based on this, a fuzzy inference algorithm of vehicle average velocity based on urgency and traffic density is put forward to predict the average speed of vehicles and the traveling time of vehicles on free travel road. The experimental results show that the simulation method based on TCPN traffic flow model and fuzzy inference can reasonably predict the travel time of vehicles in the intersection group, and compared with the Kalman filter prediction method, the proposed method has the average absolute percentage error And the root mean square error of the cumulative value decreased by 14.33% and 22.98%.