In April 2021 the US Air Force Research Laboratory awarded Lockheed Martin a 12.8 million dollar contract for the Defense Experiment in Commercial Space-Based Internet (DEUCSI) program. The DEUCSI project hopes to form a flexible, high-bandwidth, high-availability Air Force communication and data sharing capability by taking full advantage of commercial outer space-based Internet networks. The project consists of three phases, namely, using satellites and commercial demonstration terminals to establish connections between multiple Air Force sites; expanding user terminals to multiple locations and various types of platforms to extend the range of connections; and conducting specialized tests and experiments to solve special military space-based needs that cannot be met by Internet providers.
The Air Force also announced in May that the Advanced Battle Management System (ABMS) program will enter a new phase of development, moving from a focus based on testing and rapid technology development to a more traditional focus on deploying combat capabilities. The move marks the transition of ABMS into a full-fledged procurement program, moving from a largely theoretical and developmental state to one that involves specialized equipment procurement and more hands-on testing.
The Air Force Rapid Capabilities Office (RCO) has created a new capability matrix for ABMS, which includes six categories: 1. security processing; 2. connectivity; 3. data management; 4. applications; 5. sensor integration; 6. integrated effects. The Air Force plans to use more contracted tools to leverage commercial technologies, infrastructure and proven applications to get ABMS off the ground in a secure military digital network environment.
NATO is developing new cloud technologies to establish technical standards in the field and ensure interoperability among Member States. The current cloud technology project that has attracted much attention is the Firefly system, developed by French company Thales. The system will deploy NATO’s first deployable scenario-level defence cloud capability and enable its own forces to receive, analyse and transmit data between static headquarters and in real time across theatres of operation. Firefly uses an all-in-one system architecture, including application management, IT networking and security, and hence it represents a holistic approach to deployable command and control resources for the Atlantic Alliance.
Firefly is designed to provide command and control services to NATO response forces and enable collaboration between static and deployed users in support of major joint operations (MJO) or smaller joint operations (SJO). The Firefly system will provide eight deployable communication and information points of presence (DPOPs) to provide communication services with NATO command and deployed force applications and information services. Firefly will integrate and interact with existing NATO information and communication systems and provide countries and partners with Federated Mission Networking (FMN) connectivity for operations, missions and exercises so as to communicate effectively. Specific Firefly services include: communication services, infrastructure services, business support services, and staging and deployment environments.
French company Thales has been selected by the European Defence Agency (EDA) to lead the Softanet project on network scheduling and orchestration technologies. Softanet will provide insights into the use of the latest virtualization technologies in communication networks. This is an important step in preparing for the evolution of deployable tactical networks and the adoption of programmable network technologies, software-defined networking (SDN) models, and 5G.
Softanet is the first project in Europe to focus on network programmability and orchestration technologies for defence applications. Softanet will evaluate the contribution of virtualization-based network programmability and quantify the operational benefits in terms of ease of deployment, network infrastructure responsiveness, and efficient resource utilization. The three-year project will be implemented in three phases: the first six months will define a deployable network architecture based on virtualized network technologies; the subsequent eighteen months will focus on conducting research on resilience, security, and orchestration technologies; and, finally, one year will be devoted to validating and testing the system.
With the promotion of strategic competition between the great powers as the main direction, the US military pays increasing attention to comparing information in the cognitive domain, with the aim of strengthening situational awareness by enhancing public information collection, developing identification and detection technologies to protect information security and public opinion, as well as exerting influence to combat the strategic adversaries’ warfare will in view of being determined in the task, so as to carry out intelligence superiority operations and achieve the goal of “defeating the enemy without fighting.”
In August 2021 the US Defense Counterintelligence and Security Agency (DCSA) issued a request for information (RFI) for tools that can automatically search social media and other public websites: posts, actions, and interactions, to search for information regarding insider threat investigations. The tool must meet all federal and US Department of Defense’s technical requirements for access and use of government systems, and must be designed to automatically obtain open source electronic information and enable the DoD Insider Threat Management and Analysis Center investigators to search databases by name.
The tool’s retrieval results are expected to include photos, text, and actions – such as likes or retweets – taken online by “key perpetrators” without requiring DITMAC analysts to visit social media websites. Retweets are messages of no more than 140 characters in length, whose text replicates that of another message with the addition of the author’s name and any brief commentary, sent to a website via instant messenger, email or mobile phone.
The required tools must meet seven criteria: 1. be able of scanning the Internet based on known “key perpetrators”; 2. be able of performing highly accurate identity resolution based on initially limited datasets; 3. be able of scanning not only the Internet text, but also photographs and videos of images related to “key perpetrators” and “behaviours of interest”; 4. be able of providing screenshots of relevant material and be able to look more broadly at surrounding information to ensure that appropriate context is captured; 5. conduct single, ongoing inspections of “known perpetrators” at no less than weekly intervals for the period during which individual cases remain open; 6. be able to meet all Department of Defense (DoD) and Federal information technology standards to ensure the DCSA use and capability of the network; 7. be able to access all data without creating fake user accounts or establishing links to known “key perpetrators”.
In May 2021 the US Army announced it had developed a method to detect deepfakes, which could lead to the development of advanced military technology to help soldiers quickly detect and identify threats related to the aforementioned issue. Deepfakes are synthetic media in which a person in an existing image or video is replaced with someone else’s likeness. Although the act of faking content is not new, deepfakes leverage powerful machine learning and artificial intelligence techniques to manipulate or generate visual and audio content with a high potential to deceive. The main machine learning methods used to create deepfakes are based on deep learning and involve training generative neural network architectures, such as autoencoders or generative adversarial networks (GANs). The aim of this deepfake research effort is to develop a lightweight, low-cost, high-performance facial biometric recognition technology, resulting in an innovative technology solution called DefakeHop. DefakeHop’s performance is significantly ahead of the current state of the art in the industry and its key innovation is a theoretical and mathematical framework called Continuous Subspace Learning (SSL). SSL is a completely new mathematical framework for neural network architectures developed from signal transformation theory, which is entirely different from traditional methods and provides a new signal representation and process involving multiple cascading transformation matrices. SSL is a complete unsupervised data-driven framework, and provides completely new tools for image processing and for understanding tasks such as biometric facial recognition.
Giancarlo Elia Valori