基于人工神经网络的运行车速与道路安全性关系

Artificial Neural Network-Based Relation of Operating Speed and Road Safety

通过实测中国高速公路车辆运行车速并分析其道路事故特征,提出了使用运行车速以及相邻路段运行车速差预测道路安全性的概念,并运用反向传输网络(BP网络)较强的非线性适应能力,采用多层BP网络,建立了基于人工神经网络的运行车速与道路安全性关系模型.使用训练后模型测试部分实测路段安全性,取得了较好的结果,为高速公路的道路安全性评价提供了一种较为可行的方法.

The concept of predicting road safety condition through the speed difference between adjacent sections is advanced on the basis of the measurement of freeway operating speed and the analysis of accident character in China. A multiplayer back-propagation network with robust nonlinear adaptability is adopted to establish a model on the relation between operating speed and traffic safety based on Artificial Neural Network （ANN）. The tested ANN model proves feasible and effective in appraising the safety level of road section. Such a model will serve as a reliable and practicable method to evaluate freeway safety.