基于ML-II的小子样复杂系统多源信息融合方法

Method of multiple sources information fusion for complex system with small sample test based on ML-II theory

复杂系统可靠性信息的主要特点是小子样现场试验数据和多种可利用的多源先验信息。在利用Bayes理论进行小子样可靠性评定过程中,为了尽可能少做现场试验,必须充分利用各种先验信息,获取合理的先验分布。提出了一种基于第二类极大似然估计原理(ML-II)的多源信息融合方法,以某鱼雷自导系统的作用距离为例进行仿真分析,验证了该方法的合理性和有效性。

The main characteristics of reliability information of complex system are small sample field test data and various available multiple sources prior information. For small sample reliability assessment by Bayes theory, in order to do as little as possible field test and obtain reasonable prior distribution, all kinds of prior information must be made full use. The method of multiple sources information fusion for complex system with small sample test based on the Maximum Likelihood esti-mate of the second kind（ML-II）theory is brought forward and a simulation analysis is carried out in allusion to torpedo homing system’s effect distance. The results of simulation show that the fusion method is reasonable and effective.