D-S证据理论在多传感器数据融合中的应用

近年来，许多领域都在进行多传感器数据融合技术的研究。多传感器数据的属性融合有很多算法，最常用的算法是贝叶斯决策检验法，国际上已提出将证据理论用于数据融合，但在这方面的理论基础还不完善。本文研究了证据理论在多传感器数据融合中的应用。Ｄｅｍｐｓｔｅｒ－Ｓｈａｆｅｒ方法是对Ｂａｙｅｓ决策检验法的推广，证据理论比概率论满足更弱的公理系统，并且在区分不确定与不知及精确反映证据收集过程等方面显示了很大的灵活性。文中阐述了Ｄ－Ｓ证据理论的数学性质，给出了可信度公理及Ｄ－Ｓ综合规则，并进行了计算机仿真实验，实验结果说明这种判决方法非常实用，用于数据融合算法非常有效

Implementation of D-S Evidential Theory in Multisensor Data Fusion System

In recent years, numerous multisensor data fusion systems have been developed for wide applications. There are many algorithms in multisensor data attribute fusion. This paper describes the main features of the evidential combination algorithm implemented in our research. In the Bayesian approach, this theory supports the representation of uncertain information and provides a technique for combining it. The D S technique does not require prior probabilities nor does it need to know the capability of each source. The technique actually focuses on the probability of a collection of points belonging to the sample space, whereas the classical probability theory is interested in the probability of the individual points. A digital simulation has been done to demonstrate the capability of the attribute fusion algorithm.