A framework for context change detection and management in probabilistic models for context in fusion

摘要

In a prior paper, a probabilistic model for using context in fusion was developed. It was shown that context-based fusioncould be represented by a Bayesian probabilistic model that contains situation and context data, as well as conditionalprobabilities for the random variables. In the same paper, a conceptual model of an adaptive real-time contextmanagement system was proposed to monitor fusion performance, and select the appropriate context in order to improvefusion performance. This paper represents an extension of the above paper by developing frameworks for an adaptivegeneral real-time context management, using predictive analytics and modeling, with application to optimize thetracking performance of an airborne platform.