Radioactive Anomaly Detection and Identification Algorithm Suite (RADIAS) for Enhanced Radiological Search

Physical Sciences Inc. (PSI) proposes to develop an advanced machine learning (ML) algorithm to detect threat-based anomalies in gamma-ray spectra in real-time. If a network of distributed R/N sensors is employed, the algorithms will also be capable of tracking such anomalies through the network. The Radiation Anomaly Detector (RAD) will be packaged with PSI’s award winning Poisson Clutter Split (PCS) algorithm, which represents the state-of-the-art capability for real-time (1Hz) isotope identification, and PSI’s Advanced Learning-Enabled Radioactive Threat Search (ALERTS) algorithm, which is a supervised ML algorithm for isotope detection and identification. RAD will provide a complementary detection channel to PCS and ALERTS by enhancing sensitivity to anomalous or highly shielded potential threat sources. The combined Radioactive Anomaly Detection and Identification Algorithm Suite (RADIAS) will significantly improve the performance of current and future sensor systems used by US Armed Forces and Law Enforcement Agencies in the search for radioactive threats, especially Special Nuclear Material (SNM). RADIAS will be customized for integration with NGA and DoD networked sensor systems and integrated with PSI’s other commercial radiological/nuclear products such as Mobile Urban Radiation Search (MURS) and PERM-Mobile.