| 基于座椅人机参数的列车座椅舒适度评价方法 |
| |
| 引用本文： | 王金,支锦亦,向泽锐,李然,徐笑非,闫磊,徐刚.基于座椅人机参数的列车座椅舒适度评价方法[J].西南交通大学学报,2019,54(6):1342-1348. |
| |
| 作者姓名： | 王金 支锦亦 向泽锐 李然 徐笑非 闫磊 徐刚 |
| |
| 基金项目： | 国家重点研发计划资助项目（2017YFB1201103-9，2017YFB1201103-12） |
| |
| 摘 要： | 为了降低高速列车舒适度调查的成本，避免进行大量的问卷调查和统计分析，对高速列车座椅静态舒适度评价方法进行了研究. 首先，通过确定高速列车座椅舒适度评价指标和指标权重，获得座椅舒适度计算方法；其次，利用BP神经网络构建以高速列车座椅8项人机几何参数为输入、以座椅舒适度评价为输出的座椅静态舒适度评价模型；最后，进行实例研究，对构建的神经网络评价模型进行训练和验证，并对神经网络进行权值和阈值的提取，构建神经网络的数学表达公式. 研究结果表明:当神经网络为1个隐含层、13个节点时，训练达到误差均值2.13 × 10?3、均方误差6.091 × 10?6的理想效果，且不存在过拟合现象；利用CHR2的一、二等座椅人机几何参数测量数据及对应舒适度评价对该网络进行验证，验证显示一等座的神经网络预测值跟实际评价值误差为3.07%，二等座评价误差为1.42%，该网络模型预测精度较高，且优于多元回归模型预测的结果.
|
| 关 键 词： | 高速列车 座椅舒适度 神经网络 评价模型 |
| 收稿时间： | 2018-06-13 |
Evaluation Method of Seat Comfort for High-Speed Trains Based on Seat Ergonomic Parameters |
| |
| WANG Jin,ZHI Jinyi,XIANG Zerui,LI Ran,XU Xiaofei,YAN Lei,XU Gang.Evaluation Method of Seat Comfort for High-Speed Trains Based on Seat Ergonomic Parameters[J].Journal of Southwest Jiaotong University,2019,54(6):1342-1348. |
| |
| Authors: | WANG Jin ZHI Jinyi XIANG Zerui LI Ran XU Xiaofei YAN Lei XU Gang |
| |
| Abstract: | To reduce the cost in the comfort evaluation of high-speed trains and avoid strenuous questionnaire investigation and statistical analysis, the static comfort evaluation method of high speed train seats is studied. Firstly, the seat comfort calculation method is derived by determining the comfort evaluation index and index weight for high-speed train seats. Secondly, the BP neural network is used to construct a static seat comfort evaluation model, which takes the 8 ergonomic parameters of the high-speed train seat as the input and the seat comfort evaluation as the output. Finally, a case study is carried out to train and verify the constructed neural network evaluation model, and the weights and thresholds of the neural network are extracted to construct a mathematical expression of the neural network. The results show that when the neural network has one hidden layer and 13 nodes, the training achieves the desirable results with the mean error of 2.13×10?3 and mean square error of 6.091×10?6, and there is no over fitting. The network is verified by the real ergonomic data of the first-class and second-class seats in CHR2 and the corresponding comfort evaluations. The error of first-class seats between the predicted value of neural network and the actual one is 3.07%, and the error of second-class seats is 1.42%, demonstrating that the network model has high prediction accuracy and is superior to the multiple regression model. |
| |
| Keywords: | |
|
| 点击此处可从《西南交通大学学报》浏览原始摘要信息 |
|
点击此处可从《西南交通大学学报》下载全文 |
|
