Improving Radioactive Material Localization by Leveraging Cyber-Security Model Optimizations

One of the principal uses physical-space sensors in public safety applications is detection unsafe conditions (e.g., release poisonous gases, weapons airports, tainted food). However, current methods these are often costly, slow to use, and can be inaccurate complex, changing, or new environments. In this paper, we explore how machine learning used successfully cyber domains, such as malware detection, leveraged substantially enhance physical space detection. We focus on one important exemplar application-the localization radioactive materials. show that ML-based approaches significantly exceed traditional table-based predicting angular direction. Moreover, developed models expanded include approximations distance material (a critical dimension reference tables practice do not capture). With four eight detector arrays, collect counts gamma-rays features for a suite localize material. seven unique scenarios via simulation frameworks frequently radiation with experiments using laboratory observe our approach outperform standard method, reducing error by 37% reliably within 2.4%. way, advances cyber-detection provide substantial opportunities enhancing beyond.