Abstract: Based on BP neural network and RBF neural network, this paper formed a model of urban ring road traffic flow dynamic real-time forecast, and it makes predictions and analysis for the real traffic flow data of the 3er ring road in Beijing to test the validity of the model. The result of the experiment shows that this model has high accuracy. The real-time feature of the algorithm can satisfy the practical forecast system.

Key words: traffic flow forecast, BP neural network, RBF neural network, urban traffic

摘要： 基于BP神经网络和RBF神经网络构建了城市环路交通流动态实时预测模型，并对北京市三环路的实测交通流数据进行了预测分析。实证结果表明，该模型具有较高的精确度，算法的动态实时性也满足实际预测系统的要求，具有较好的实用性。

关键词: 交通流预测, BP神经网络, RBF神经网络, 道路交通