Support Intelligence and Fusion Technology to Ecologically Represent and Reason for Air Cargo Scanners (SIFTER-ACS)

Air cargo security is critical to minimizing the overall risk to national security. However, screening air cargo with current state-of-the-art computed tomography (CT) imaging is a complex and challenging task for screeners. To address current complexities and challenges, Charles River Analytics proposes to design and demonstrate Supportive Intelligence and Fusion Technology to Ecologically Represent and Reason for Air Cargo Screeners (SIFTER-ACS). SIFTER-ACS is an intuitive, ecological human-machine interface that represents cargo risk using probabilistic reasoning to assess manifest data and scanning outputs providing screeners with decision-critical and context-rich information to support scan strategy and interpretation. First, we will perform a cognitive analysis of the air cargo work domain to develop human-system interaction and work-support requirements for the SIFTER-ACS solution. Second, we will employ dual-step Natural Language Processing (NLP) of text extraction and semantically map manifest contents to a common data schema to automatically extract decision-critical information from the cargo manifest. Third, we will leverage the dual node decision wheels (DNDW) architecture for data fusion and response management to prevent unnecessary skid breakdown and delays. Finally, we will employ novel ecological interface design methods to drive the design of an intuitive and useful human-machine interface (HMI). This approach will represent high-level analytic conclusions about cargo manifest data to inform scanning strategy and help air cargo screeners reason about ambiguous scans. Combined, the SIFTER-ACS will provide screeners with decision-critical information and enriched context to support scan interpretation while simultaneously promoting efficiency and safety.