System and method for symbol decoding in high frequency (hf) communication channels

Patent metadata

Field	Value
Publication number	US-12432007-B2
Application number	US-202217963894-A
Country	US
Kind code	B2
Filing date	Oct 11, 2022
Priority date	Oct 11, 2022
Publication date	Sep 30, 2025
Grant date	Sep 30, 2025

Abstract

One embodiment provides a method and a system for reconstructing symbols transmitted over a high frequency (HF) communication channel. During operation, the system can receive, at a receiver, a radio frequency (RF) signal carrying an input data frame and transmitted over the HF communication channel. The input data frame includes a number of known symbols followed by a number of unknown symbols. The system can determine a set of channel parameters associated with the HF communication channel based on the received RF signal and the known symbols and reconstruct, using a machine-learning technique, the unknown symbols based on the determined channel parameters and the received RF signal.

First claim

What is claimed is:

1. A method for reconstructing symbols transmitted over a high frequency (HF) communication channel, the method comprising:
receiving, at a receiver, a radio frequency (RF) signal carrying an input data frame and transmitted over the HF communication channel, wherein the input data frame comprises a number of known symbols followed by a number of unknown symbols;
determining a set of channel parameters associated with the HF communication channel based on the received RF signal and the known symbols; and
reconstructing, using a machine-learning technique, the unknown symbols based on the determined channel parameters and the received RF signal, wherein the reconstruction comprises:
determining, based on the received RF signal, a current mode of the HF communication channel;
selecting, from a plurality of decoder models which are previously trained offline, a decoder model corresponding to the current mode of the HF communication channel; and
reconstructing the unknown symbols based on the channel parameters and the selected decoder model.

2. The method of claim 1 , wherein determining the channel parameters comprises performing a gradient-based optimization operation based on a loss function indicating a difference between a response of the HF communication channel to the known symbols and a response of a model of the HF communication channel with the determined channel parameters to the known symbols.

3. The method of claim 1 , wherein reconstructing the unknown symbols further comprises:
optimizing parameters of a decoder according to corresponding parameters of the selected previously trained decoder model; and
reconstructing the unknown symbols using the decoder with the optimized parameters.

4. The method of claim 3 , wherein optimizing the parameters of the decoder comprises performing a gradient-based optimization operation based on a cross-entropy loss function.

5. The method of claim 4 , wherein the parameters of the decoder are optimized jointly with parameters of a corresponding encoder that encodes the input data frame.

6. The method of claim 1 , wherein determining the current mode of the HF communication channel comprises applying a cluster-analysis technique on the received signal.

7. The method of claim 1 , wherein determining the current mode of the HF communication channel comprises:
decoding, in parallel, the received RE signal using the trained decoder models to obtain a plurality of decoded data frames; and
comparing bit error rates (BERs) of the plurality of decoded data frames based on the known symbols in the input data frame.

8. The method of claim 1 , wherein reconstructing the unknown symbols comprises solving an integer programing problem to directly predict the unknown symbols.

9. The method of claim 8 , wherein solving the integer programing problem comprises:
solving a relaxed integer programing problem by allowing the predicted unknown symbols to have continuous values; and
rounding up the predicted unknown symbols with continuous values to nearest integers.

10. A computer system for reconstructing symbols transmitted over a high frequency (HF) communication channel, the computer system comprising:
a processor; and
a storage device coupled to the processor and storing instructions, which when executed by the processor cause the processor to perform a method, the method comprising:
determining a set of channel parameters associated with the HF communication channel based on a radio frequency (RF) signal received over the HF communication channel, wherein the received RF signal is encoded based on an input data frame comprising a number of known symbols followed by a number of unknown symbols; and
reconstructing, using a machine-learning technique, the unknown symbols based on the determined channel parameters and the received RF signal, wherein the reconstruction comprises:
determining, based on the received RF signal, a current mode of the HF communication channel;
selecting, from a plurality of decoder models which are previously trained offline, a decoder model corresponding to the current mode of the HF communication channel; and
reconstructing the unknown symbols based on the channel parameters and the selected decoder model.

11. The computer system of claim 10 , wherein determining the channel parameters comprises performing a gradient-based optimization operation based on a loss function indicating a difference between a response of the HF communication channel to the known symbols and a response of a model of the HF communication channel with the determined channel parameters to the known symbols.

12. The computer system of claim 10 , wherein reconstructing the unknown symbols further comprises:
optimizing parameters of a decoder according to corresponding parameters of the selected previously trained decoder model; and
reconstructing the unknown symbols using the decoder with the optimized parameters.

13. The computer system of claim 12 , wherein optimizing the parameters of the decoder comprises performing a gradient-based optimization operation based on a cross-entropy loss function.

14. The computer system of claim 13 , wherein the parameters of the decoder are optimized jointly with parameters of a corresponding encoder that encodes the input data frame.

15. The computer system of claim 10 , wherein determining the current mode of the HF communication channel comprises applying a cluster-analysis technique on the received signal.

16. The computer system of claim 10 , wherein determining the current mode of the HF communication channel comprises:
decoding, in parallel, the received RF signal using the trained decoder models to obtain a plurality of decoded data frames; and
comparing bit error rates (BERs) of the plurality of decoded data frames based on the known symbols in the input data frame.

17. The computer system of claim 10 , wherein reconstructing the unknown symbols comprises solving an integer programing problem to directly predict the unknown symbols.

18. The computer system of claim 17 , wherein solving the integer programing problem comprises:
solving a relaxed integer programing problem by allowing the predicted unknown symbols to have continuous values; and
rounding up the predicted unknown symbols with continuous values to nearest integers.

Assignees

• Palo Alto Res Ct Inc
• Xerox Corp

Inventors

• Matei Ion
• De Kleer Johan

Classifications

• G06N3/045

  Combinations of networks · CPC title

• H04L1/0036

  arrangements specific to the receiver · CPC title

• G06N3/088

  Non-supervised learning, e.g. competitive learning · CPC title

• H04L25/0254

  using neural network algorithms · CPC title

• G06N3/084

  Backpropagation, e.g. using gradient descent · CPC title