Skip to main content
ARS Home » Research » Research Project #442870

Research Project: Bio-hardened Environments: Integrating AI-based Biothreat Detection Networks into Situational Awareness and Rapid Response Systems

Location: Research Programs

Project Number: 3022-32000-018-013-S
Project Type: Non-Assistance Cooperative Agreement

Start Date: Sep 15, 2022
End Date: Apr 30, 2027

The objective of this project is to demonstrate a bio-hardened environment – an area where the introduction of biothreats can be rapidly detected and countered. This approach borrows strategies from the cyber domain: (1) fast digital sensing using software-upgradable threat signatures to detect key biothreats in real-time, (2) area-wide anomaly detection to identify novel biothreat intrusions beyond known threat lists, and (3) tight integration into command and control systems to deploy countermeasures rapidly and effectively. This project will be conducted in the Korean Peninsula in collaboration with United States Forces Korea (USFK) utilizing Microsoft’s Premonition technologies for AI-based biomonitoring. It is supported by the U.S. Department of Agriculture National Bio and Agro-Defense Facility (USDA NBAF). Lessons learned will be used to protect critical infrastructure and economic sectors from biothreats – in domestic and international contexts.

This project will focus on layering Microsoft Premonition’s Biological Weather Stations (BWTs) into existing situational awareness and rapid response systems to form an AI-based biothreat detection network. BWTs are deployed into environments and provide three novel capabilities: (1) low-cost digital bio-sensing to identify specific disease vectors providing real-time biological weather, (2) over-the-air updates of AI-based biothreat recognizers ensuring agility, (3) robotic bio-sampling to acquire specific environmental bio-samples for downstream pathogen analysis, based on programmable criteria. A network of BWTs provides an additional layer of capabilities: (1) unsupervised construction of area-wide baselines of nominal environments by observing many different species populations, (2) notification and robotic bio-sampling of anomalies that deviate from baselines, which could be expected in novel biothreat intrusions or in rapid degradation of environments, and (3) network-wide missions that optimize use of consumables and minimize personnel exposure to high-risk environments. In summary, this will create a sensing and sampling layer that can respond to a wide range of biological phenomena in complex environments. This effort will also demonstrate data integration. Microsoft Premonition systems are built using modern IoT communications protocols, cloud-scale data lake architectures, and user experiences for detailed tracking of samples from environment through laboratory workflows. This will provide a testbed to integrate multi-modal, remotely sensed, and real-time data into established situational awareness and rapid response systems. Guidance from stakeholders, such as USDA, DTRA, DoD GEIS, and NAMRU, will help refine data interfaces that integrate BWT capabilities into major and existing infrastructure and missions. Taken together, the resulting bio-hardening blueprints will: (1) define the scope and characteristics of ideal environments to strategically bio-harden, (2) characterize the core infrastructure and cyber capabilities needed to bio-harden those environments, and (3) determine key sensing infrastructure and software adaptation points to efficiently integrate future biosensing modalities and analytics capabilities.