Detecting radioactivity is inherently uncertain. In a typical passive detection scenario (e.g. scanning cargo coming off a ship), a large number of factors need to be taken into
account, all of which can be subject to varying levels of uncertainty. For example, sensor count rates can vary, cargo can contain benign sources of radiation (e.g. bananas or tiles) that could mask other proscribed materials, the shipping manifest might not be clear, etc. Ultimately, this uncertainty tends to be handled on a subjective basis. Individual operators take what are often ad-hoc decisions. This means that there can be a large degree of variation in terms of how operators handle detection events such as alarms. In this project we attempt to apply a statistical multi-faceted decision support technique called Evidential Reasoning (ER) to provide a more systematic means by which to address this uncertainty. ER represents a decision problem as a hierarchy of factors, ranging from the top overarching question, down to the low-level atomic factors. It enables an operator (or indeed a sensor depending on the context) to supply an opinion on each of the lowest-level sub-factors, and automatically amalgamates this information to provide a high-level outcome. Importantly, this outcome makes any doubt or uncertainty explicit and quantifiable.