基于BP神经网络的高速公路动态交通流预测

以高速公路交通流预测为研究对象,建立了基于BP神经网络的参数动态修复交通流预测模型。以高速公路宏观动态交通流模型为原型,利用分段辨识法分析了高速公路交通流特性。对BP神经网络层数和神经元的确定,以及转移函数的优化选择进行了深入研究,并给出了基于BP神经网络交通流预测模型的建模方法。对西宝(西安-宝鸡)高速公路交通流实时数据进行了采集、建模和仿真。通过仿真结果与实际结果比较,验证了该模型具有较高的可信度。

Expressway Dynamic Traffic Flow Prediction Based on BP Neural Network

This paper takes expressway traffic flow prediction as object of study,and sets up a traffic flow prediction model based on dynamic parameter correction of BP neural network.It takes the expressway macro dynamic traffic flow model as prototype and analyzes the characteristics of expressway traffic flow with "Segment Identification Method".It also carries out detail researches on determination of BP neural network levels and neural elements,as well as optimization of transition functions,and it gives modeling method based on BP neural network traffic flow prediction model.It carries out real data collection from traffic flow at Xi'an-Baoji Expressway,modeling and emulation.Through comparison emulation result and actual result,this model is proved to be highly reliable.