基于动态概率神经网络的公路交通事件检测算法

Detection Algorithm for Expressway Traffic Events Based on Dynamic Probability Neural Network

从公路隧道事件及其自身特点出发,依据人工神经网络这一强大的无模型模式分类器,采用人工神经网络算法建立动态概率神经网络(DPNN)事件检测模型。模型采用的输入指标不是单独的交通流参数,而是几个参数的组合,充分利用交通数据中包含的交通事件信息,大大提高算法的性能。

Starting from expressway tunnel events and their features and based on artificial neural network--a powerful non - model mode classifier, this paper adopts artificial neural network algorithm to establish dynamic probability neural network （DPNN） event detecting model. The input indices used by the model are not single traffic flow parameters, but the combination of a few parameters, thus the traffic event information included in traffic data can be adequately utilized to greatly improve the performance of the algorithm.