...
  • 主题
高级搜索  >
历史搜索 清空历史
首页> 外文期刊>Accident Analysis & Prevention >A Bayesian Neural Network approach to estimating the Energy Equivalent Speed
【24h】

A Bayesian Neural Network approach to estimating the Energy Equivalent Speed

机译：贝叶斯神经网络方法估算能量等效速度

获取原文
获取原文并翻译 | 示例
           
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

To reduce the number and the gravity of accidents, it is necessary to analyse and reconstruct them. Accident modelling requires the modelling of the impact which in turn requires the estimation of the deformation energy. There are several tools available to evaluate the deformation energy absorbed by a vehicle during an impact. However, there is a growing demand for more precise and more powerful tools. In this work, we express the deformation energy absorbed by a vehicle during a crash as a function of the Energy Equivalent Speed (EES). The latter is a difficult parameter to estimate because the structural response of the vehicle during an impact depends on parameters concerning the vehicle, but also parameters concerning the impact. The objective of our work is to design a model to estimate the EES by using an original approach combining Bayesian and Neural Network approaches. Both of these tools are complementary and offer significant advantages, such as the guarantee of finding the optimal model and the implementation of error bars on the computed output. In this paper, we present the procedure for implementing this Bayesian Neural Network approach and the results obtained for the modelling of the EES: our model is able to estimate the EES of the car with a mean error of 1.34 m s~(-1). Furthermore, we built a sensitivity analysis to study the relevance of model's inputs.
机译：为了减少事故的数量和严重性，有必要进行分析和重建。事故建模需要对冲击进行建模，进而需要估计变形能。有几种工具可用来评估车辆在撞击过程中吸收的变形能量。但是，对更精确，更强大的工具的需求不断增长。在这项工作中，我们将车辆在碰撞过程中吸收的变形能量表示为能量等效速度（EES）的函数。后者是难以估计的参数，因为在碰撞过程中车辆的结构响应不仅取决于与车辆有关的参数，而且取决于与冲击有关的参数。我们工作的目的是设计一种模型，通过结合贝叶斯和神经网络方法的原始方法来估计EES。这两种工具是互补的，并提供了显着的优势，例如保证找到最佳模型，并在计算出的输出上实现误差线。在本文中，我们介绍了实现这种贝叶斯神经网络方法的过程以及对EES建模的结果：我们的模型能够估计汽车的EES，平均误差为1.34 m s〜（-1）。此外，我们建立了敏感性分析来研究模型输入的相关性。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号