Analysis and Comparison of Probability Transformations for Fusing Sensors with Uncertain Detection Performance

Abstract

In a recent paper by Davey, Legg and El-Mahassni a way of fusing sensors with uncertain performance was outlined using the Transferable Belief Model (TBM) theory. It showed that if the target prior was uncertain, then the resulting fused mass was also uncertain. That is, some belief mass was assigned to the case that the presence or absence of a target was unknown. Various methods have been proposed to transform an uncertain belief function into a probability mass. This paper analyses the relationship between an important subset of these methods and compares the resulting probability masses with those obtained via Bayesian methods using random priors.