Dependable and generic high - level information fusion - methods and algorithms for uncertainty management Technical report HS - IKI - TR - 07 - 003

The main goal of information fusion can be seen as exploiting diversities in information to improve decision making. The research field of information fusion can be divided into two parts: low-level information fusion and high-level information fusion. Most of the research so far, has concerned the lower levels, e.g., signal processing and multi-sensor data fusion, while high-level information fusion, e.g., clustering of entities, has been relatively uncharted. High-level information fusion aims at providing decision support (human or automatic) concerning situations. A crucial issue for decision making based on such support is trust, defined as “accepted dependence”, where dependence or dependability is an overall term for other concepts, e.g., reliability. Dependability requirements in high-level information fusion refer to properties of belief measures and hypotheses regarding situations. Even though meeting such requirements is considered to be a precondition for trust in fusion-based decision-making; research in high-level information fusion that addresses this issue is scarce. Since most of the research in high-level information fusion relate to defense applications, another important issue is to generalize existing terminology, methods, and algorithms, in order to allow for researchers in other domains to more easily adopt such results. In this report, it is argued that more research is needed for these issues and a set of research questions for future research is presented.