基于改进的Boosting算法的高速公路交通事件检测

Freeway Traffic Incident Detection Based on Improved Boosting Method

提出一种基于Boosting BP神经网络的交通事件检测方法.以上下游的流量和占有率作为特征,BP神经网络作为分类器进行交通事件的自动分类与检测.为了进一步提高神经网络的泛化能力,采用一种调整权值分布,限制权重扩张的改进的Boosting方法,分类器以加权投票方式进行分类决策.实验结果表明该交通事件检测算法是有效的.

A novel method is proposed for freeway traffic incidents detection based on boosting BP neural network. The features of flow and occupancy rate are extracted from traffic incidents. Then BP neural network is used to classify the traffic incidents. In order to improve the precision of the BP neural network for traffic incidents detection, a weight-adjusting-based method is proposed to improve Boosting method. The experiment results show that the proposed method is effective in detection the traffic incidents.