高速公路交通流宏观模型的一种分类辨识算法

Identification approach to macro model of traffic flow in freeway using parameter classification

高速公路交通流模型是一个高阶非线性时变系统 ,这使得该模型的辨识问题成为一个非常困难的问题。简要介绍了模型及其中各参数的含义 ,在对模型参数加以分析、讨论的基础上将其分类并分别采用径向基函数 (RBF)神经网络和最小二乘法对模型的参数进行分类辨识 ,成功地解决了该模型辨识的工程化问题。通过与传统的复合形法的辨识结果进行比较 ,该方法的辨识精度和速度均明显提高 。

The macro model of traffic flow in freeway is a high order, nonlinear and time variant system which makes the problem of its identification become very difficult. A simple introduction of the module and the meaning of its parameters was given out. The parameters of the model were classified and identified by using radial basis function (RBF) neural net (NN) and Least Square method. The result of the simulation was given which indicates that the training speed and the identification accuracy of the method are better than that of the complex method.