TASCS: Tactical AI Spectrum Classifier System for Agile Spectrum Superiority

The use of the electromagnetic spectrum (EMS) to carry out offensive, defensive and supportive operations is increasingly being shaped by new digital capabilities to gain superiority in today’s evolving battlefield. These new capabilities pose a growing threat to Unites States airborne platforms’ ability to operate and defeat adversaries. The United States needs to ensure it always retains airspace dominance, by maintaining superior EMS capabilities, which can be afforded through technological innovation. Using Machine Learning (ML) to automatically detect and classify adversarial and unintentional interference can be a key enabler in accelerating the response to interference and the adoption of automated countermeasures in weapon systems. However, the substantial variations and complexities of missions, surroundings, and types of systems, limitations of small real-world data sets, as well as the need for continuous deployable capability pose unique challenges to conventional ML methods. Qylur’s ‘Social Network of Intelligent Machines’ (SNIM®) is an innovative core AI technology and platform that maintains the long-term performance of Artificial Intelligent/Machine Learning (AI/ML) and Deep Learning (DL) models that work with complex data under changing conditions. SNIM® AI performs automatic sensor data ingestion, model accuracy monitoring, optimization retraining, and tactical re-deployment of models to weapon systems to ensure continued superiority.In this phase II effort, we propose to adapt Qylur’s SNIM® AI platform and further extend it to include domain-specific AI-based interference detection components, both software and hardware, to provide the DAF and the DoD with a new Tactical AI-based Spectrum Classifier System (TASCS). This objective of this R&D effort is to provide the DAF and DoD with a new capability to expedite analysis processes and enable the new autonomous system responses for countermeasures to protect EMS compromised systems.