Edge Computing for AI/ML based in forward deployed Cell Phones and associated equipment

RT&L FOCUS AREA(S): Artificial Intelligence/ Machine Learning; Cybersecurity; Network Command, Control and Communications TECHNOLOGY AREA(S): Battlespace; Electronics; Information Systems; Sensors OBJECTIVE: Explore novel approaches to increase sense making for deployed forces using equipment they already carry and use. This will enable AI/ML Sense Making closer to the user with fewer delays, and reduced communications requirements, leading to a reduction of the human factors burden and physical burden by augmenting capabilities in their existing equipment. DESCRIPTION: Develop, or demonstrate existing capabilities to create virtual processing networks using the increasing computational capabilities built in to commercially available cell phones. The ability to do secure processing close to the end user presents a rich resource for moving advanced and distributed processing AI/ML closer and closer to the end user. This will decrease the communications burden or bandwidth required while decreasing the time from potential detection to final action. There is a movement to use inexpensive commercial equipment to enable operators in the field. The Android Tactical Assault Kit and it variants are now standard equipment for deploying forces. More specific to DTRA the Joint CTTSO DTRA Joint Operation Center/Tactical Assault Kit project provides sharing of imagery, planning tools, planning fragments, standard formatted request and report messages, language translation, and many additional operator developed capabilities. The old Galaxy S10 phones used with JOCTAK include dual CPU processors, 16 megapixel cameras, and at least 128 GB of ram1. The newer S21 version includes the faster dual processors, up to 512 GB of ram, a 40 MP Camera, dual encryption chips, and 5G capability2. Each generation provides more capability without development cost to the end user. The goal of this SBIR/SSTR topic is to explore harvesting and applying computational capabilities available in the current and next generation of cell phones used in DTRA Tactical Assault Kits, the Joint Operations Center Tactical Assault Kit (JOCTAK) and with the AI/ML capabilities under research in other parts of CX. The emerging application and use of TAK/ATAK with the emergence of Civilian Tactical Assault Kit or the Team Awareness Kit is indicative of the commercial application of this research. PHASE I: PHASE I: Shall execute an Analysis of Alternatives (AofA) and produce a Proof of Concept (PoC) implementation and architecture for creating secure virtual processing networks across commercially available cell phone used in the TAK/JOCTAK3 projects. PHASE II: Shall mature and fully integrate the architecture for the virtual network application with the continuous integration, continuous delivery (CICD) automated build process, located at the DI2E collaboration portal (www.di2e.net). PHASE III DUAL USE APPLICATIONS: All the deliverables of this topic shall be either Open Source (OS) or UNLIMITED RIGHTS and the DI2E collaboration portal shall be used for increased transparency, reduced cost and schedule, and to increase performance. SBIR/SSTR Rights are not permitted because the target transition architecture does not accept software with SBIR/SSTR Rights. Having ensured the widest availability across DoD and the government enable much wider applicaiton of the capability. REFERENCES: 1. https://www.samsung.com/global/galasxy-s10-/specs/; 2. https://www.samsung.com/global/galasxy-s21-ultra-5g/specs/; 3. https://www.civtak.org/tag/socom;