Evolving Towards a Reference Architecture for Data Fusion

Data and Information Fusion technologies have evolved and matured in implementation over the last four decades and are considered a key value adding capability for a wide range of complex systems. Inclusion of data and information fusion technology in a system, however, can add to the complexities of engineering the system because data and information fusion is complex in its own regards and lack a system engineering reference architecture that can be readily employed in systems. A reference architecture is an emerging concept in systems engineering which aims to provide an authoritative resource regarding a specific subject area that guides and constraints instantiation of multiple architectural solutions within that domain. While a number of technological models and architectures are proposed in data fusion literature (e.g., the JDL Data Fusion Model), none of these models provide necessary and sufficient guidance expected of a reference architecture for deriving fusion solutions in modern complex systems. Given the advances in data and information over the last four decades, this paper first establishes the need for building a reference architecture for data and information fusion domain and then examines how the current body of knowledge and practice can be synthesized towards realizing a data fusion reference architecture.