Definition of bba and quality of fusion in C2 systems

This paper presents the results of numerical experiments devoted to examination of influence of the target attribute hypotheses definitions on quality of information fusion. The main goal of the research works presented herein was to find answers to the subsequent questions: `In the context of attribute information fusion in C2 systems, is it reasonable to extend the information scope of a sensor?´ and `Does the extension of the sensor information scope always provide tangible benefits in quality of fusion?´ In order to achieve that there have been defined two measures: decision robustness and decision bias. In the experimentation Dezert-Smarandache Theory has been used as the main fusion engine. In this paper, both Bayesian and non-Bayesian basic belief assignments have been considered.