系统工程与电子技术 ›› 2019, Vol. 41 ›› Issue (7): 1544-1550.doi: 10.3969/j.issn.1001-506X.2019.07.15

• 系统工程 • 上一篇    下一篇

基于改进型BP神经网络的导引头测高性能评估

何华锋,何耀民,徐永壮   

  1. 火箭军工程大学导弹工程学院, 陕西 西安 710025
  • 出版日期:2019-06-28 发布日期:2019-07-09

High performance evaluation of seeker measurement based on improved BP neural network

HE Huafeng,HE Yaomin,XU Yongzhuang   

  1. Department of Missile Engineering, Rocket Force University of Engineering, Xi’an 710025, China
  • Online:2019-06-28 Published:2019-07-09

摘要: 武器装备性能评估贯穿于装备发展的全寿命过程,对于装备定型与实战化运用具有重要意义。针对传统反向传播(back propagation, BP)神经网络模型易陷入局部最优、武器装备评估数据少等问题,提出了基于数据包络分析和Bootstrap法的改进型BP神经网络模型。利用数据包络分析处理原始指标得到3项优化参数,结合Bootstrap法对其进行扩充,再通过BP神经网络建立评估模型。实验表明,改进模型得到的合成孔径雷达(systemic aperture radar, SAR)导引头测高性能评估结果,其决定系数和误差系数均有明显改善。该模型不仅规避了模糊综合评判法主观性强、精确度不高等问题,同时有效解决了传统BP神经网络模型易陷入局部最优和武器装备评估数据少的两个问题。

关键词: 性能评估, 神经网络, 合成孔径雷达导引头

Abstract: The performance evaluation of weapon equipment runs through the whole life process of equipment development, which is important to equipment finalization and practical application. In view of that the traditional back propagation (BP) neural network is prone to local optimal and weapon equipment evaluation data are a small sample type, this paper proposes an improved BP neural network model based on data envelopment analysis and the Bootstrap method. Three optimized parameters are obtained by data envelopment analysis, and expanded by the Bootstrap method, and then an evaluation model is established by the BP neural network. The experimental results show that the determination coefficient and error coefficient of the systemic aperture radar seeker are improved obviously. This model not only avoids the problems of strong subjectivity and low accuracy of the fuzzy comprehensive evaluation method, but also solves two shortcomings of the traditional BP neural network model which is prone to be involved in local optimization and the lack of weapon equipment data.


Key words: performance evaluation, neural network, systemic aperture radar (SAR) seeker