MADECLEAR: Machine-Assisted Detection of Concealed Signals for Electronic Attack Response

Near-peer adversaries employ advanced communication and radar systems whose emissions are difficult to detect or uniquely identify, particularly on a tactically relevant timescale. Conventional electronic support (ES) techniques are computationally intensive and cannot reliably detect obfuscated signals in real time. These current brute-force methods will be less effective in the future high-end fight, putting US assets and personnel at risk. To counter these risks and achieve spectrum dominance in a conflict, it is essential to efficiently detect challenging low-probability intercept/low-probability of detection (LPI/LPD) RF emissions.  Further, it is necessary to do so in real time and “at the edge”, enabling direct feedback into kinetic or non-kinetic effector systems.  To address this challenging detection problem, Boston Fusion Corp. (BFC) and Raytheon propose Machine-Assisted Detection of Concealed Signals for Electronic Attack Response (MADECLEAR). MADECLEAR employs advanced AI/ML techniques to directly detect and classify LPI/LPD signals in the presence of noise and electromagnetic (EM) interference, using only direct-sampled I/Q RF data as input.