Method and apparatus for transmitting channel state information

Patent metadata

Field	Value
Publication number	US-12355522-B2
Application number	US-202318482468-A
Country	US
Kind code	B2
Filing date	Oct 6, 2023
Priority date	Oct 7, 2022
Publication date	Jul 8, 2025
Grant date	Jul 8, 2025

Abstract

An operating method of a wireless communication device includes receiving a reference signal from a base station, estimating a first channel between the wireless communication device and the base station based on the reference signal, extracting, based on a first artificial intelligence model trained to reduce a difference between the first channel estimated by the wireless communication device and a second channel estimated by the base station, a feature including grouped attributes from the first channel, quantizing the grouped attributes using a codebook that is based on a second artificial intelligence model and by generating one or more indices of the codebook for each group of the attributes, and transmitting, to the base station, a bitstream including combination information of the indices of the codebook.

First claim

What is claimed is:

1. A method of a wireless communication device, the method comprising:
receiving a reference signal from a base station;
estimating a first channel between the wireless communication device and the base station based on the reference signal;
extracting, based on a first artificial intelligence model trained to reduce a difference between the first channel estimated by the wireless communication device and a second channel estimated by the base station, a feature including grouped attributes from the estimated first channel;
quantizing the grouped attributes using a codebook that is based on a second artificial intelligence model by generating one or more indices of the codebook for each group of the attributes; and
transmitting, to the base station, a bitstream including combination information of the indices of the codebook,
wherein the codebook includes a plurality of vector codewords and a plurality of scalar codewords, and the second artificial intelligence model comprises an artificial intelligence model trained to reduce a difference between a combination of the plurality of vector codewords and the plurality of scalar codewords of the codebook and the quantized attributes.

2. The method of claim 1 , wherein the first artificial intelligence model is trained to group attributes having high correlation.

3. The method of claim 1 , wherein the first artificial intelligence model is trained based on an importance of the grouped attributes.

4. The method of claim 1 , wherein the first artificial intelligence model is trained based on a density of the grouped attributes.

5. The method of claim 4 , wherein the first artificial intelligence model is trained to select attributes that are greater than 0 from each group of the grouped attributes.

6. The method of claim 4 , further comprising selecting one or more attributes from each group of the attributes, based on the first artificial intelligence model, wherein the combination information of the indices of the codebook is based on a quantization value of the selected one or more attributes from each group of the attributes and position information of the selected one or more attributes.

7. The method of claim 1 , further comprising selecting, based on a bitmap, some groups from among the groups,
wherein the combination information of the indices of the codebook comprises information corresponding to one or more of the selected groups.

8. The method of claim 1 , wherein the bitstream is transmitted on an uplink channel.

9. The method of claim 1 , further comprising:
receiving, by the base station, the bitstream;
segmenting, by the base station, the bitstream using a fourth artificial intelligence model;
obtaining, by the base station, the quantized attributes using a codebook based on the fourth artificial intelligence model;
obtaining, by the base station, the second channel with respect to the first channel that is estimated by the wireless communication device using a third artificial intelligence model from the quantized attributes;
wherein, before the wireless communication device performs operations with the base station,
parameters of the first artificial intelligence model and the second artificial intelligence model are updated as the wireless communication device trains the first artificial intelligence model and the second artificial intelligence model, and
parameters of the third artificial intelligence model and the fourth artificial intelligence model are updated as the base station trains the third artificial intelligence model and the fourth artificial intelligence model.

10. The method of claim 9 , wherein the first artificial intelligence model and the second artificial intelligence model are trained in parallel, and
the third artificial intelligence model and the fourth artificial intelligence model are trained in parallel.

11. The method of claim 1 , wherein the reference signal comprises any one of a channel state information-reference signal (CSI-RS), a synchronization signal block (SSB), a demodulation-reference signal (DM-RS), and a tracking reference signal (TRS).

12. The method of claim 1 , wherein the first artificial intelligence model and the second artificial intelligence model comprise non-linear relationships.

13. A method of a wireless communication device comprising:
receiving a reference signal from a base station;
estimating a first channel between the wireless communication device and the base station based on the reference signal;
extracting, based on a first artificial intelligence model trained to reduce a difference between the first channel estimated by the wireless communication device and a second channel estimated by the base station, a feature including grouped attributes from the estimated first channel;
generating one or more indices of a codebook for each group of the attributes by quantizing the grouped attributes using the codebook based on a second artificial intelligence model; and
transmitting, to the base station, a bitstream including the indices of the codebook,
wherein the codebook comprises a plurality of vector codewords, wherein the second artificial intelligence model comprises an artificial intelligence model trained to reduce a difference between a combination of the plurality of vector codewords of the codebook and the quantized attributes.

14. The method of claim 13 , wherein the first artificial intelligence model is trained to group attributes having high correlation.

15. The method of claim 13 , wherein the first artificial intelligence model is trained based on an importance of the grouped attributes.

16. The method of claim 13 , wherein the first artificial intelligence model is trained based on a density of the grouped attributes.

17. The method of claim 16 , wherein the first artificial intelligence model is trained to select attributes that are greater than 0 from each group of the grouped attributes.

18. The method of claim 16 , further comprising selecting one or more attributes from each group of the attributes, based on the first artificial intelligence model,
wherein combination information of the indices of the codebook is based on a quantization value of the selected one or more attributes from each group of the attributes and position information of the selected one or more attributes.

19. The method of claim 13 , further comprising selecting, based on a bitmap, some groups from among the groups,
wherein combination information of the indices of the codebook is based on at least one of the selected groups.

20. A method of a base station comprising:
receiving, from a wireless communication device, a bitstream including information about one or more codebook indices for quantized attributes;
segmenting, by the base station, the bitstream using a second artificial intelligence model;
extracting, by the base station, the quantized attributes from the segmented bitstream using a codebook based on the second artificial intelligence model; and
obtaining, by the base station, a second channel from the quantized attributes using a first artificial intelligence model for reducing a difference between a first channel estimated by the wireless communication device and a second channel estimated by the base station,
wherein the codebook includes a plurality of vector codewords and a plurality of scalar codewords, and the second artificial intelligence model comprises an artificial intelligence model for reducing a difference between a combination of

Assignees

• Samsung Electronics Co Ltd

Inventors

• Lee Junho
• Kim Jaein
• Je Huiwon
• Joo Hyunseung

Classifications

• H04L25/0254

  using neural network algorithms · CPC title

• H04L5/0048

  Allocation of pilot signals, i.e. of signals known to the receiver (allocation of control signalling H04L5/0053; use of control signalling H04L5/0091) · CPC title

• H04B17/3913

  Predictive models, e.g. based on neural network models · CPC title

• H04L25/0224

  using sounding signals · CPC title

• H04B7/0639

  Using selective indices, e.g. of a codebook, e.g. pre-distortion matrix index [PMI] or for beam selection · CPC title