Novel AI/ML/Deep-Learning (DL) based PNT Signal Classifier (DLPC)

Effective Navigation Warfare (NAVWAR) planning in the A2/AD environment requires real-time situational awareness (SA) information about the unintended signals (jammers and spoofers) that may be present in the frequency bands of interest.  The NAVWAR planning also requires understanding of the effectiveness of Blue Force Electronic Attack (EA).In this Army SBIR program, Mayflower proposes to further develop, implement, test and demonstrate a novel Deep Learning-based Position, Navigation and Time (PNT) signal classifier (DLPC) applied to GNSS spectrum.  Unique features of the proposed DLPC technology include: (i) better classification accuracy compared to conventional classification techniques; (ii) net training optimized to require very minimal amount of real test data compared to Convolutional Neural Networks (CNN) employed in traditional machine learning; (iii) optimized to minimize memory and processing requirements compared to traditional DL techniques; (iv) ability to readily incorporate within GNSS Digital Antenna Electronics (DAE) Anti-Jam (AJ) solution in software-defined system architecture; (v) better suited for rapid classification of emerging threat; and (vi) Size, weight and power - optimized combined SA/AJ solution. The Phase II program development effort will primarily focus on the development and demonstration of DLPC prototype for classification of jammer/spoofer signals in the GNSS band in a DAE.