A CBRN detection framework using fuzzy logic

Identifying a chemical, biological, radiological, and nuclear incident (CBRN) is a challenge. Evidence and health symptoms resulting from CBRN malevolent incident overlap with other normal non malevolent human activities. However, proper fusion of symptoms and evidence can aid in drawing conclusions with a certain degree of credibility about the existence of an incident. There are two types of incidents directly observable, overt, or indirectly observable, covert, which can be detected from the symptoms and consequences. This paper describes a framework for identifying a CBRN incident from available evidence using a fuzzy belief degree distributed approach. We present two approaches for evidence fusion and aggregation; the first, two level cumulative belief degree (CBD) while the second is ordered weighted aggregation of belief degrees (OWA). The evaluation approach undertaken shows the potential value of the two techniques.