Advanced Sensor Fusion Using Low-Cost Sensors for Dramatic Physical Security Cost Reduction

Physical security is a major cost driver for nuclear sites, including a large up-front investment in perimeter intrusion detection systems (PIDS) and millions more for security forces to monitor and respond to alarms. The future of advanced, small modular, and microreactors (ASMR) depends on cost reductions to compete with other sources of electricity. Advances in sensor technology by themselves have failed to reduce the high nuisance and false alarm rates (NAR/FAR) that exacerbate monitoring costs. The cluttered environments around ASMR sites are expected to further challenge security, driving up costs while allowing adversaries to exploit vulnerabilities caused by the day-to-day activities at these locations. Management Sciences, Inc. (MSI), in SBIR Phase III R&D with Sandia National Labs (SNL), has developed and demonstrated an exciting new high-performance threat detection and tracking framework called Dynamic Motion Analytics (DMA). DMA leverages sensor fusion and a novel form of motion analytics that distinguishes intruders from benign background activity and noise, dramatically improving intruder detection performance by filtering out alarms caused by weather, wildlife, clutter, benign pedestrian and vehicle traffic, and other noise. In DoE funded pilot deployments at nuclear sites, DMA not only achieved dramatic reductions in NAR/FAR but also showed exciting potential to eliminate or reduce fixed PIDS investments for ASMR applications. We now propose extensions to DMA that will enhance its excellent intruder detection performance with layers and methods for verification and validation of intrusion events in real time. Our enhanced DMA product vision is an integrated security concept where human response forces are required less often, and those forces will be better prepared with more detailed information about the intrusion event and its actors and goals. Evolving security systems with the emerging capabilities of AI automation will advance the state of the art from mere detection and response to situation awareness, enabling a single central alarm station (CAS) to monitor multiple sites, continuously providing up-to-date detailed information about the type of intruders, their numbers, points and times of entry, and predicted paths and locations through the site, enabling rapid and optimal security force responses at substantially lower cost over current methods. Our proposed R&D effort includes the research, design, prototyping, and demonstration of the multi-layer DMA capability. In collaboration with SNL, our DoE SBIR sponsor, and experienced industry and academic subject matter experts, we will architect DMA layers that facilitate automated real-time threat assessment and response that drive deep cost reductions for ASMR physical security. The multi-layer DMA system will lower the infrastructure and labor costs of security at high-traffic sites such as hospitals, airports, schools, manufacturing facilities, businesses, and fixed or temporary US military installations. The system will help site designers assess the security impacts and costs of tradeoff decisions early during site design while improving performance and reducing operational costs when deployed, reducing the total cost of ASMR physical security.