An Evidential Reasoning Rule Considering Parameter Uncertainty

Jie Wang,Zhi-Jie Zhou,Chang-Hua Hu,Xiao-Xia Han,Shuai-Wen Tang,Peng-Yun Ning

DOI: https://doi.org/10.1109/TAES.2021.3115076

IF: 3.491

2022-01-01

IEEE Transactions on Aerospace and Electronic Systems

Abstract:This article aims to propose an evidential reasoning (ER) rule considering the parameter uncertainty. As the essential parameters, the evidence weight and reliability make the ER rule constitute a generalized reasoning framework. Theoretically, the weight is affected by subjective cognition, while the reliability mainly reflects objective variation. However, most of the recent researches have focused on the quantitative calculation methods, which make the differences in the property of the two parameters ignored. In this article, a relatively different idea from previous studies is provided, in which multisource uncertainty of parameters is fully considered. On the one hand, the weight is profiled by interval variable with lower and upper bounds. On the other hand, the reliability is modeled by random variable with probability distribution function. Then, a unified inference model for evidence aggregation is developed based on the inference process of the ER rule. In addition, some basic properties of the model are clearly presented to illustrate the rationality of parameter uncertainty. Finally, a practical example is given to show the potential applications of the proposed model.