Abstract: In this paper，by analyzing the freeway macroscopic dynamic traffic flow model，Back Propagation（BP） neural network model is presented aiming at nonlinear and time-variant system of freeway traffic flow.Then，the BP neural network is trained by using traffic flow data from a section of freeway and the network model parameters can be obtained.Finally，the model is simulated under MATLAB for verifying its effectiveness and the simulation results are compared with the original model ones.It shows that the proposed method can accurately describe the real behavior of freeway traffic flow and can adapt well to the changes of traffic condition.

Key words: expressway, traffic flow, modeling, BP neural network

摘要： 通过对高速公路宏观动态交通流模型的分析，针对高速公路交通系统的非线性时变特点，应用BP神经网络建立了高速公路宏观动态交通流模型。并利用一段高速公路的交通流数据对BP神经网络进行训练，得到网络参数。最后，为了验证BP网络模型的有效性，在MATLAB环境中对模型进行了仿真，并将仿真结果与原始模型的结果进行了比较。结果表明，该方法能较准确地描述高速公路交通流的真实行为，并且能够适应交通状况的变化。

关键词: 高速公路, 交通流, 建模, BP神经网络