高速公路动态交通系统的神经网络预测控制

NEURAL NETWORK PREDICTIVE CONTROL FOR THE DYNAMIC TRAFFIC SYSTEM OF THE EXPRESSWAY

提出一种高速公路动态交通系统的神经网络预测控制方法．根据高速公路的实际背景导出其非线性动态模型，利用神经网络及辨识的串并联结构来建立模型的映射，在此基础上给出控制变量带有上下界约束的预测控制算法．根据实测数据进行了仿真研究，结果表明算法是有效的．因而为解决高速公路的动态控制问题提供了一种新的方法．

This paper proposes a kind of neural network predictive control method for the dynamic traffic system of the expressway. It analyses the nonlinear structure of the expressway and sets up its nonlinear mapping by using the serialparallel identifying structure of the neural network. On the basis, a neural network predictive control algorithm for the problem in which the input variables possess the upper and lower constraints is given. The simulation study is completed according to the real measuring data and results show that the algorithm is effective and a new method is provided for solving the dynamic control problem of the expressway.