US 12,413,336 B2
Methods and apparatus for decoding received uplink transmissions using log-likelihood ratio (LLR) optimization
Sabih Guzelgoz, San Jose, CA (US); and Hong Jik Kim, San Jose, CA (US)
Assigned to Marvell Asia Pte, Ltd, Singapore (SG)
Filed by Marvell Asia Pte, Ltd., Singapore (SG)
Filed on Dec. 4, 2022, as Appl. No. 18/074,506.
Application 18/074,506 is a continuation of application No. 17/173,041, filed on Feb. 10, 2021, granted, now 11,558,147.
Claims priority of provisional application 62/975,080, filed on Feb. 11, 2020.
Prior Publication US 2023/0120502 A1, Apr. 20, 2023
This patent is subject to a terminal disclaimer.
Int. Cl. H04L 1/00 (2006.01); G06N 5/04 (2023.01); G06N 20/00 (2019.01)
CPC H04L 1/0054 (2013.01) [G06N 5/04 (2013.01); G06N 20/00 (2019.01)] 25 Claims
OG exemplary drawing
 
1. An apparatus for processing data transmitted via a communication network, comprising:
a soft-demapper configured to generate log-likelihood ratios (LLR) values via soft-demapping resource elements in accordance with soft-demapping parameters;
a decoder coupled to the soft-demapper and configured to generate decoded data based on the LLR values;
a performance metric circuit coupled to the decoder and identifying a performance metric from the decoded data; and
a processor coupled to the performance metric circuit and configured to adjust one or more soft-demapping parameters in response to the performance metric and a target performance in accordance with an operation of a machine learning algorithm, wherein the processor is able to update the soft-demapping parameters based on a selected LLR bit-width value via performance of the machine learning algorithm.