基于Adaboost集成RBF神经网络的高速公路事件检测

提出一种基于Adaboost集成RBF神经网络的高速公路事件检测方法。首先对高速公路事件检测原理进行分析，进行了相关的参数选择，确定了RBF神经网络的结构，然后采用改进的Adaboost方法集成RBF神经网络进行高速公路事件检测并给出了事件检测算法的步骤，最后进行了仿真实验，实验结果表明，该方法可以明显提高RBF神经网络性能（高检测率、低误报率），且具有较强的泛化能力，适宜高速公路事件检测。

Freeway incident detection based on Adaboost RBF neural network

A method based on Adaboost RBF neural network is proposed for freeway incident detection in this paper.By analyz- ing the principles of freeway incident detection and choosing the relative parameters,the topology of neural network is fixed.Then an algorithm utilizing Adaboost integrating RBF neural network for freeway incident detection and steps of its processing are pre- sented.Finally,a simulation experiment is given.The result of the experiment shows that this algorithm can highly enhance the performance of RBF neural network (with high detection rate and low false alarm rate) and also has a good performance of generalization,which assesses the effectiveness of its application to freeway incident detection.