RadioNuclide Threat Tracking System (RNTTS)

The integration of radiological and nuclear (RN) sensors with commercial video management systems (VMSs) would greatly enhance DHS/CWMD's ability to fulfill its mission. To demonstrate the value of integrating RN sensors with existing VMSs, DA proposes the RadioNuclide Threat Tracking System (RNTTS). RNTTS interfaces with existing RN sensors and uses cutting-edge machine learning (ML) algorithms to identify and track RN carriers in the video streams within VMSs. During Phase II DA will improve the accuracy of RNTTS and demonstrate that the solution can scale to use cases where: 1) many potential RN carriers are present in complex environments and 2) many security cameras are available to continuously track RN carriers within the VMS. During Phase II several strategies will be utilized to improve the accuracy and scalability of the Phase I RNTTS solution. From a hardware perspective, RNTTS will utilize state-of-the-art GPUs for computing. From an algorithm perspective, the RNTTS algorithm will be upgraded to accommodate simultaneous tracking and identification of many RN carriers. From a software perspective, key portions of the underlying RNTTS algorithms will be parallelized to support real-time operations on commercial VMSs. During Phase II RNTTS will be demonstrated, tested, and evaluated (DT&E) at two locations with two commercial RN sensors and two commercial VMS vendors. DT&E efforts during Phase II will enable DA to assess RNTTS' performance at real-world deployments, which is critical for commercialization. The completed Phase II prototype will enable drop-in integration of RNTTS for the select VMS vendors and RN sensors.