基于深度度量学习的导弹气动系数预测
作者:
作者单位:

1.西南科技大学计算机科学与技术学院,绵阳 621000;2.中国空气动力研究与发展中心计算空气动力研究所,绵阳 621000;3.中国航发四川燃气涡轮研究院,绵阳 621000

作者简介:

通讯作者:

杨春明,男,副教授,E-mail: yangchunming@swust.edu.cn。

中图分类号:

V211.24

基金项目:

四川省科技厅重点研发项目(2021YFG0031);先进航空动力创新工作站项目(HKCX2022-01-022)。


Missile Aerodynamic Coefficient Prediction Through Deep Metric Learning
Author:
Affiliation:

1.School of Computer Science and Technology, Southwest University of Science and Technology, Mianyang 621000, China;2.Computational Aerodynamic Research Institute, China Aerodynamics Research and Development Center, Mianyang 621000, China;3.AECC Sichuan Gas Turbine Establishment, Mianyang 621000, China

Fund Project:

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
  • |
  • 文章评论
    摘要:

    传统多输出深度神经网络在导弹气动性能系数预测任务中,通常采用均方误差(Mean square error, MSE)和平均绝对误差(Mean absolute error, MAE)来训练网络,但在小样本及无物理方程约束的情况下,MSE与MAE对导弹性能系数之间的约束和不同导弹样本之间的区分就会降低。针对该问题,提出一种基于深度度量学习的K最近邻大边距损失函数(K-nearest neighbor large margin, KNNLM),它通过边距约束将大差异输出样本推开,拉近相近输出样本,以此来解决样本及样本间的约束区分问题。以导弹气动外形及工况参数作为输入,4种气动系数作为输出,在反向传播神经网络(Backpropagation neural network, BPNN)和多任务学习神经网络(Multi-task learning neural network, MTLNN)中分别采用MSE、MAE、KNNLM进行实验对比,实验结果表明:KNNLM在BPNN和MTLNN中的精度相比于MSE和MAE最大能够提升14.44%和16.35%,最少提升3.72%。KNNLM能够在少样本及无物理知识约束的情况下,能更好地对导弹样本进行约束区分,使深度神经网络模型的预测精度更高,且鲁棒性更强。

    Abstract:

    Mean square error (MSE) and mean absolute error (MAE) are usually used to train traditional multi-output deep neural networks in missile aerodynamic coefficient prediction. However, in the case of small sample size and no physical equation constraint, the constraint between MSE and MAE on missile performance coefficient and the distinction between different missile samples will be reduced. A K nearest neighbor large margin (KNNLM) loss function based on deep metric learning is proposed. The method uses the margin constraint to push the output samples with large differences away, and close the similar output samples. Taking the aerodynamic shape and working condition parameters of the missile as input and four aerodynamic coefficients as output, MSE, MAE and KNNLM are used for experimental comparison in backpropagation neural network (BPNN) and multi-task neural network (MTLNN). The experimental results show that compared with MSE and MAE, KNNLM can improve the accuracy by 14.44% and 16.35% at most, and 3.72% at least in BPNN and MTLNN. And the KNNLM can better distinguish the missile samples in the case of fewer samples and no physical knowledge constraint, so that the prediction accuracy of the deep neural network model is higher and the robustness is stronger.

    参考文献
    相似文献
    引证文献
引用本文

刘林,杨春明,蔺佳哲,向宏辉.基于深度度量学习的导弹气动系数预测[J].南京航空航天大学学报,2024,56(5):950-959

复制
分享
文章指标
  • 点击次数:
  • 下载次数:
  • HTML阅读次数:
  • 引用次数:
历史
  • 收稿日期:2023-05-08
  • 最后修改日期:2023-10-08
  • 录用日期:
  • 在线发布日期: 2024-11-08
  • 出版日期:
文章二维码
您是第位访问者
网站版权 © 南京航空航天大学学报
技术支持:北京勤云科技发展有限公司