基于神经网络技术的危险货物运输驾驶员安全评价系统研究

Research of Driver Safety Evaluation System for Dangerous Goods Transportation Based on Neural Network Technology

分析引发危险货物运输事故的产生原因,由驾驶员引起的事故占较大的比例。文章采用以评价促进管理的方法来管理驾驶员,预防事故发生。首先对行业企业进行调研,构建考核驾驶员的安全评价指标体系,建立了四个一级指标,二十六个二级指标。构建了驾驶员三层神经网络评价模型,Levenberg-Marqut神经网络优化算法在文中使用,并进行了100步以上的训练,达到了精度要求。采用Mat Lab模拟仿真训练算法,训练结果表明,该方法适应性好。

Analysis of the causes of accidents in the course of the transport of dangerous goods,accidents caused by the driver accounted for a large proportion. The paper adopts the method of promoting management by evaluation to manage the drivers to prevent the accident. The paper constructs the index system for the driver's safety evaluation,which includes 4 first-level indexes and 26 secondary indexes,after the research of enterprises. The paper constructs three-layer neural network of driver evaluation model, uses the Levenberg-Marqut neural network optimization algorithm,and practices more than 100 steps to achieve the accuracy requirement. The results of training show that the good adaptability to the method of Mat Lab simulation training algorithm.