基于迭代学习辨识算法的非线性宏观交通流参数优化  

Nonlinear macroscopic traffic flow parameters optimizationbased on iterative learning identification algorithm

作　　者：胡小勇 张建军[2] 邓志刚 杨云晖 HU Xiaoyong;ZHANG Jianjun;DENG Zhigang;YANG Yunhui(Xinyu Highway Development Center,Xinyu 338099,China;Key Laboratory of Integrated Transportation Big Data Application Technology,Beijing Jiaotong University,Beijing 100044,China;Xinyu Highway Survey and Design Institute,Xinyu 338099,China)

机构地区：[1]新余市公路事业发展中心,江西新余338099 [2]北京交通大学综合交通运输大数据应用技术交通运输行业重点实验室,北京100044 [3]新余公路勘察设计院,江西新余338099

出　　处：《现代电子技术》2023年第13期183-186,共4页Modern Electronics Technique

基　　金：江西省交通运输厅科技项目(2021H0023)。

摘　　要：考虑到迭代学习辨识算法具有的独特优势,设计了一种基于迭代学习辨识算法的非线性宏观交通流参数优化方法。通过一个城市道路网络的模拟实验表明,该方法能够有效地识别时变量多参数的动态系统。研究结果表明:采用VISSIM的路段评估函数对各个路段的密度和车速进行采集,可方便对迭代式识别方法评估结果进行比较。随着迭代次数的增多,网络各个路段排队车错误值逐渐降低,且保持不变。对网络交通系统的最大错误对比,更好地体现了迭代识别的正确性。道路网的非线性宏观流量模型与道路模拟实验结果总体上与道路交通流量的实际改变相一致,验证了该方法在道路网络中的非线性大流量模型的识别性能。Considering the unique advantages of iterative learning identification algorithm,a method of nonlinear macroscopic traffic flow parameter optimization based on iterative learning identification algorithm is designed.The simulation experiments of an urban road network show that this method can effectively identify the dynamic system with time variable and multiple parameters.The results show that:the road segment evaluation function of VISSIM is used to collect the density and speed of each road segment,so as to facilitate the comparison of the evaluation results of the iterative identification method.With the increase of iteration times,the error values of queuing vehicles on each section of the network gradually decrease and remain unchanged.The maximum error comparison of the network transportation system can better reflect the correctness of iterative identification.Although the nonlinear macroscopic flow model of road network and the results of road simulation test are generally consistent with the actual change of road traffic flow,the identification performance of the method in the nonlinear large flow model of the road network is verified.

关 键 词：交通流 非线性 拥堵现象 迭代学习 参数辨识 承载能力 交叉口信号 

分 类 号：TN911.1-34[电子电信—通信与信息系统] TP13[电子电信—信息与通信工程]