Principles of a unified framework for heterogeneous information fusion and assessment

This paper presents the ongoing development of Futhane, a framework intended to support heterogeneous information fusion while providing mechanisms for performances assessment. The overall approach is based on the use of functional components, offering basic processing functionalities to be combined in a variety of fusion strategies suitable for situation awareness. The framework is adaptable thanks to both its open architecture and the use of various external resources. Therefore, users are allowed to implement new functional components, to develop alternative fusion strategies and to customize the processing chains by using appropriate maps, taxonomies or ontologies. The paper outlines the architecture of the framework, its main functional components. Uncertainty issues are considered in addition to functional capabilities, and specific means are provided to assess the outcome of components. Thus, the outcome can be assessed by following a step-by-step process, which makes possible to understand the impact of each component within the fusion chain. The paper also illustrates several studies to be conducted thanks to the Futhane framework in order to highlight the value of fusing hard and soft information for various applications.