A data association framework for general information fusion

We extend the concept of data association as defined in the context of target tracking to general fusion and call it general data association. The main differences are that general data association encompasses heterogeneous observations (e.g. from social media text, and UAV images), multiple observations from the same data, multiple assignments of observations to entities, and association measures tailored for both observations and environment entities. We propose a framework to encapsulate different aspects of general data association, apply it to an intelligence analysis scenario, and point to some directions for future research. Our general data association framework should serve as a stepping stone for discussion and further exploration of the detected challenges.