ABSTRACT
This chapter focuses on the identification function and on comparisons of Bayesian and Dempster–Shafer methods for identity estimation in data fusion. A fusion process improves the perception of the environment thanks to the multiple interpretations provided by different sensors. However, any information processing is based on information content. In the case of Bayesian mass allocation, Probability and Evidence theories are in full concordance. It is when non-Bayesian allocation is provided that differences between the two theories can be highlighted. The chapter illustrates and explains these aspects. Each sensor delivers masses on union of elementary hypotheses. To adapt the information to probabilistic combination, it is necessary to share the masses on each union between all elementary hypotheses that compose it. An electronic support measure (ESM) receiver is used to detect, measure, and segregate streams of incoming pulses coming from RADAR. This gives rise to radar interceptions. These radar interceptions have to be compared to emissions characteristics stored in a database.
