An information fusion framework for threat assessment

Modern enterprises are becoming increasingly sensitive to the potential destructive power of small groups or individuals with malicious intent. In response, significant investments are being made in developing a means to assess the likelihood of certain threats to their enterprises. Threat assessment needs are typically focused in very specific application areas where current processes rely heavily on human analysis to both combine any available data and draw conclusions about the probability of a threat. A generic approach to threat assessment is proposed, including a threat taxonomy and decision-level information fusion framework, that provides a computational means for merging multi-modal data for the purpose of assessing the presence of a threat. The framework is designed for flexibility, and intentionally accounts for the accuracy of each data source, given the environmental conditions, in order to manage the uncertainty associated with any acquired data. The taxonomy and information fusion framework is described, and discussed in the context of real-world applications such as shipping container security and cyber security.