Adapting commercial technologies to deliver the Modular Attributable Sensor System (MASS), an array of AI-enabled sensor nodes interoperable with the Unified Network

OUSD (R&E) CRITICAL TECHNOLOGY AREA(S): Ingrated Network Systems of Systems OBJECTIVE: Develop and validate a software tool compatible with various hardware systems that will allow military end users to analyze, store, and share photo and video data. The system will allow military customers to tap real-time data sources from IP-based CCTV, mobile ISR systems and unattended sensors (e.g. camera traps, small tactical grids) in a unified interface. Sample use cases include physical security, fire/smoke surveillance, firearm detection, and wildlife surveys on training lands. The goal of the project is to develop a specific use case, contributing to the Army Network modernization priority. DESCRIPTION: Today, large Army installations rely on sparse, labor-intensive patrols for security, training operations, compliance, and other routine operational tasks. As a result, these areas suffer from high operational costs, slow response times, and frequent disruptions to training activities, directly impacting military readiness. Persistent, AI-based surveillance can modernize many of these routine tasks when deployed at scale across these areas to free Army personnel to focus on the mission, reduce operational costs, and improve response times for critical security, compliance, and operational events (e.g. unknown vehicle intrusion in restricted areas, vandalism of government property, disease spreading among a population of a federally listed species). Successful solutions should provide Army Network compatibility, have scalability for large quantities of data, offer web-based command and control interface, and allow forward and backward integration of various sensor systems through zero-trust APIs. The solution should be adaptable across various use cases in line with Army installation of the future priorities. Sensor pods that incorporate renewable technology (e.g. solar power) have the potential to contribute to the Energy Independence and Security Act of 2007 priorities, while offering improved installation resiliency. PHASE I: Perform lab testing and customer discovery of a system that will allow military end users to analyze video and photo data. The system should be compatible with the Army Network and provide operational benefits (e.g., time savings) in a specific use case. Quantify the accuracy of detection for different events. Determine how artificial intelligence algorithms can be incorporated into DoD and commercial operations. The false positive rate and false negative rate should reach < 90%. PHASE II: Develop a new system or adapt a commercial product that will allow military end users to use AI approaches to categorize imagery and other related data. Test the system in real-world use with a partner installation identified by the TPOC. The false positive rate and false negative rate should reach < 95%. Reduce the time for photo categorization by 50+% and improve data availability. PHASE III DUAL USE APPLICATIONS: The project has broad applicability across all military branches/installations and private sector. The outcome of the project will be a state of art system reflecting learnings from military, commercial and open-source communities. The benefits of the project will include (1) Time savings through the use of automation and collaboration tools, (2) Faster detection speed for operational trends and adaptive REFERENCES: 1. U.S. Army Engineer Research and Development Center. (n.d.). Virtual Testbed for Installation Mission Effectiveness Archives . Power of ERDC Podcast. https://poweroferdcpodcast.org/tag/virtual-testbed-for-installation-mission-effectiveness/ 2. Microsoft. (2020, November 22). Microsoft Rocket for Live Video Analytics. Microsoft Research. https://www.microsoft.com/en-us/research/project/live-video-analytics/ 3. What Is Amazon Rekognition? (1:42). (n.d.). Amazon Web Services, Inc. https://aws.amazon.com/rekognition/ 4. Army TM 11-6675-379-10c. "Operator's Manual for Instrument Set, Reconnaissance and Surveying (ENFIRE) AN/TKQ-5 (NSN: 6675-01-559-6558)," Army Publishing Directorate, 2009. KEYWORDS: artificial intelligence, unified network, security surveillance, sensor agnostic, persistent power management, renewable energy, climate resilience, natural and cultural resource management