US 12,223,443 B1
Learning approximate estimation networks for communication channel state information
Timothy James O'Shea, Arlington, VA (US); Kiran Karra, Arlington, VA (US); and T. Charles Clancy, Arlington, VA (US)
Assigned to Virginia Tech Intellectual Properties, Inc., Blacksburg, VA (US)
Filed by Virginia Tech Intellectual Properties, Inc., Blacksburg, VA (US)
Filed on Jul. 6, 2023, as Appl. No. 18/218,855.
Application 17/732,683 is a division of application No. 16/017,904, filed on Jun. 25, 2018, granted, now 11,334,807, issued on May 17, 2022.
Application 18/218,855 is a continuation of application No. 17/732,683, filed on Apr. 29, 2022, granted, now 11,699,086.
Claims priority of provisional application 62/523,861, filed on Jun. 23, 2017.
Int. Cl. H04B 17/309 (2015.01); G06N 3/02 (2006.01); G06N 5/046 (2023.01); G06N 20/00 (2019.01); H04B 17/391 (2015.01)
CPC G06N 5/046 (2013.01) [G06N 3/02 (2013.01); G06N 20/00 (2019.01); H04B 17/309 (2015.01); H04B 17/3912 (2015.01)] 20 Claims
OG exemplary drawing
 
1. A method of deploying one or more machine-learning estimation network to a communications system, the method comprising:
receiving, by at least one receiver, a radio frequency (RF) signal transmitted through a communication channel;
generating, by a machine-learning estimation network, estimated channel state information of the communication channel through which the RF signal was transmitted, wherein the machine-learning estimation network is trained based on at least one channel effect of the communication channel; and
generating, using the estimated channel state information of the communication channel and the RF signal transmitted through the communication channel, a recovered RF signal by correcting an offset of the RF signal using the estimated channel state information.