US 11,916,515 B2
Demodulation of non-return to zero (NRZ) or other signals using multiple-input artificial intelligence/machine learning (AI/ML) model
Jacob M. Miller, State College, PA (US); and H. Brown Cribbs, III, Arlington, VA (US)
Assigned to Raytheon Company, Tewksbury, MA (US)
Filed by Raytheon Company, Waltham, MA (US)
Filed on Feb. 14, 2022, as Appl. No. 17/670,814.
Prior Publication US 2023/0261613 A1, Aug. 17, 2023
Int. Cl. H03D 9/00 (2006.01); G06N 3/04 (2023.01); G06N 20/00 (2019.01); H04L 27/26 (2006.01)
CPC H03D 9/00 (2013.01) [G06N 3/04 (2013.01); G06N 20/00 (2019.01); H04L 27/2649 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method comprising:
obtaining pairs of in-phase (I) and quadrature (Q) samples associated with a signal to be demodulated;
providing a set of the I/Q pairs to a trained artificial intelligence/machine learning (AI/ML) model, the set of the I/Q pairs including an I/Q pair associated with a symbol being demodulated and at least one I/Q pair associated with at least one prior symbol that has been demodulated; and
using the trained AI/ML model to generate a symbol estimate for the symbol based on the set of the I/Q pairs, wherein the symbol estimate represents a portion of data that is encoded in the signal.