航空学报 > 2016, Vol. 37 Issue (10): 3110-3120   doi: 10.7527/S1000-6893.2016.0151

直升机火控系统精度敏感性分析的BNSobol法

贺楚超, 高晓光   

  1. 西北工业大学 电子信息学院, 西安 710129
  • 收稿日期:2015-10-22 修回日期:2016-02-15 出版日期:2016-10-15 发布日期:2016-05-27
  • 通讯作者: 高晓光,Tel.:029-88431256 E-mail:cxg2012@nwpu.edu.cn E-mail:cxg2012@nwpu.edu.cn
  • 作者简介:贺楚超 男,博士研究生。主要研究方向:贝叶斯网络学习与应用。Tel:029-85269057 E-mail:xomrssh@163.com;高晓光 女,博士,教授,博士生导师。主要研究方向:智能决策,复杂系统建模与效能分析。Tel:029-88431256 E-mail:cxg2012@nwpu.edu.cn
  • 基金资助:

    国家自然科学基金(61573285)

Sensitivity analysis on accuracy of helicopter fire control system: A BNSobol method

HE Chuchao, GAO Xiaoguang   

  1. School of Electronics and Information, Northwestern Polytechnical University, Xi'an 710129, China
  • Received:2015-10-22 Revised:2016-02-15 Online:2016-10-15 Published:2016-05-27
  • Supported by:

    National Natural Science Foundation of China (61573285)

摘要:

针对现有研究大都只分析了单误差源对直升机火控系统精度影响的现状,提出进行直升机火控系统精度的敏感性分析。根据传统敏感性分析法在数据量不充分或无法提供精确模型时存在的局限性,对基于贝叶斯网络和Sobol指数相结合的敏感性分析新方法(BNSobol法)进行了研究,并与传统Sobol法对比分析。结果表明BNSobol法在满足精度要求的同时还减小了对数据量的需求,它的分析不依赖于模型且可以确定出满足约束条件的具体误差源取值区间。该方法可为直升机火控系统设计研究时如何分配各模块误差大小从而提高总体效能提供参考和理论支持。

关键词: 直升机, 精度评估, 敏感性分析, 火控系统, 贝叶斯网络

Abstract:

Pointing at the status that most of the current research only analyzed the effect that a single error source works on the helicopter fire control system, we propose that the sensitivity analysis on accuracy of the helicopter fire control system should be executed. To achieve more accurate sensitivity analysis with sparse data and imprecise model, a new sensitivity analysis method named BNSobol method is proposed based on Bayesian network and Sobol index theory. The proposed BNSobol method is experimentally compared with classical Sobol method. Experiments reveal that the proposed method can reduce the amount of required analysis data, meanwhile subject to certain accuracy. The analysis does not rely on the model and it can present concrete error source's value region under certain constraints. The BNSobol method provides theory support and alternative method for error allocation to each part, which further raises total efficacy of the helicopter fire control system.

Key words: helicopter, accuracy analysis, sensitivity analysis, fire control system, Bayesian network

中图分类号: