摘要
针对雷达导引头机电结构组成复杂、性能指标测试数据信息利用率不足、使用传统基于数据驱动的状态预测方法精度不高的问题,借鉴相关向量机(relevance vector machine,RVM)和Dempster-Shafer(D-S)证据理论,提出了一种基于证据融合和改进局域RVM的状态预测方法。首先,对标准RVM回归模型进行改进,通过构建方差高斯核函数(variance Gauss kernel function,VGKF)来提高核函数的全局性能和泛化能力;然后通过借鉴混沌序列局域预测法中邻近点个数的选取方法,利用Hannan-Quinn(H-Q)准则对训练空间预测嵌入维数进行优化,避免了主观选取的盲目性,完成了改进局域相关向量机模型(local relevance vector machine,LRVM)的构建;最后,利用具有近似退化规律的同源装备测试数据对LRVM进行了改进,通过D-S证据理论对两种模型的预测结果进行了融合,建立了联合局域相关向量机(united local relevance vector machine,U-LRVM)模型。通过对导引头相关参数的实例预测,验证了该方法的可行性和优越性。
For the problems of complex mechanical and electrical structure of radar seeker,insufficient utilization rate of test data information and low accuracy of the traditional state prediction method based on data driven,based on the relevance vector machine(RVM)and the Dempster-Shafer(D-S)evidence theory,a state prediction method is proposed based on evidence fusion and improved local RVM(LRVM).Firstly,to improve the standard RVM regression model,variance Gauss kernel function(VGKF)is constructed to improve the global performance and generalization ability of kernel function.Then,by using the chaotic sequence local prediction selection method of the number of neighboring points of the law,the training space prediction of embedding dimension is optimized by Hannan-Quinn(H-Q)criterion.The blindness of subjective selection is avoided and the improved LRVM model is constructed.Finally,the LRVM is improved by using the homology equipment test data with approximate degradation law.Based on the D-S evidence theory,the prediction results of the two models are fused,and a united LRVM(U-LRVM)model is established.The feasibility and superiority of the proposed method are verified by an example of the correlation parameters of the seeker.
出处
《系统工程与电子技术》
EI
CSCD
北大核心
2018年第1期9-16,共8页
Systems Engineering and Electronics
基金
国家自然科学基金(51605487)
山东省自然科学基金(ZR2016FQ03)资助课题
关键词
相关向量机
D-S证据理论
方差高斯核函数
局域预测法
relevance vector machine(RVM)
Dempster-Shafer(D-S)evidence theory
variance Gauss kernel function(VGKF)
local prediction method
