BioSIGMA: Sensor Expansion and Turbulence Modeling

Physical Sciences Inc. (PSI) proposes the enhancement and maturation of BioSIGMA, a persistent early-warning capability for the detection and identification of aerosolized biological warfare agents. The BioSIGMA technology utilizes PSI’s DisperseNET, an advanced suite of algorithms that correlate spatio-temporal sensor responses with real-time physical dispersion modeling to estimate relevant source terms, including release location and plume extent. BioSIGMA uses this dispersion modeling to power real-time simulations where bio point-sensor and anemometer measurements are then fused within a Bayesian Sequential Estimation framework to reduce alarms from clutter and confusers by an order of magnitude. This capability has been demonstrated to achieve a 1 in 30-day false alarm rate at the network-level. This proposed sequential Phase II effort will update DisperseNET with a machine-learning based computational fluid dynamics solver that is capable of reproducing high-quality outputs on the order of milliseconds, compared to tens of minutes for traditional software. Additionally, the proposed effort will enhance BioSIGMA by implementing simulation support for mobile sensor platforms, and developing a data-driven normalcy modeling approach for BWA confusers in cluttered urban environ­ments. Finally, PSI will update BioSIGMA for enhanced correlation of species identification outputs from emerging BWA sensing technologies.