5G signals detection using neural network

Patent metadata

Field	Value
Publication number	US-11700585-B2
Application number	US-202117246800-A
Country	US
Kind code	B2
Filing date	May 3, 2021
Priority date	Sep 16, 2019
Publication date	Jul 11, 2023
Grant date	Jul 11, 2023

Abstract

A communication system that includes neural network system configured to learn from training examples of radio frequency (RF) signals, and circuitry configured to label a first set of fifth generation (5G) RF signals originated from a first type of source and a second set of 5G RF signals originated from a second type of source. At least one RF impairment is added randomly to each labelled example of the first set of 5G RF signals and the second set of 5G RF signals, wherein randomization of artificially added RF impairment to each labeled example corresponds to addition of different RF impairments randomly to different labeled examples. The neural network system is trained with a plurality of labelled examples. Each labelled example includes an artificially added RF impairment. The circuitry uses the trained neural network system to detect an input 5G RF signal having a new RF impairment.

First claim

What is claimed is:

1. A communication system, comprising:
a neural network system configured to learn from training examples of radio frequency (RF) signals; and
circuitry configured to:
label a first set of fifth generation (5G) radio frequency (RF) signals originated from a first type of source and a second set of 5G RF signals originated from a second type of source;
artificially add at least one RF impairment randomly to each labelled example of the first set of 5G RF signals and the second set of 5G RF signals, wherein randomization of artificially added RF impairment to each labeled example corresponds to addition of different RF impairments randomly to different labeled examples;
train the neural network system with a plurality of labelled examples, wherein each labelled example includes the artificially added RF impairment; and
execute a plurality of functionalities of a 5G modem based on the trained neural network system.

2. The communication system according to claim 1 , wherein the first set of 5G RF signals are software application-generated 5G RF signals with known labels, and wherein the second set of 5G RF signals are 5G RF signals captured over-the-air (OTA) from one or more 5G radio access nodes (RANs).

3. The communication system according to claim 1 , wherein the circuitry is further configured to decode synchronization signals from 5G new radio (NR) synchronization signal (SS) burst that includes a plurality of synchronization signal blocks (SSBs), based on the trained neural network system, wherein the 5G NR SS burst is transmitted by use of one or more directional beams by a base station and captured by the communication system to decode the synchronization signals based on the trained neural network system for time-frequency synchronization.

4. The communication system according to claim 3 , wherein the circuitry is further configured to extract master information block (MIB) based on the trained neural network system, wherein the MIB is broadcast system information periodically transmitted by the base station.

5. The communication system according to claim 3 , wherein the circuitry is further configured to extract beam identifier (ID) of a beam of 5G RF signal having a highest Received Signal Strength Indicator (RSSI) among a plurality of RSSI associated with a plurality of beams of 5G RF signals, based on the trained neural network system.

6. The communication system according to claim 1 , wherein the circuitry is further configured to utilize the trained neural network system to detect an input 5G RF signal having a new RF impairment.

7. The communication system according to claim 6 , wherein the circuitry is further configured to extract physical cell identifier (ID) of a serving base station from the input 5G RF signal based on the trained neural network system.

8. The communication system according to claim 6 , wherein the circuitry is further configured to:
decode a physical downlink shared channel (PDSCH) used for downlink data transmission from the input 5G RF signal; and
decode a physical uplink shared channel (PUSCH) used for uplink data transmission from the input 5G RF signal based on the trained neural network system.

9. The communication system according to claim 6 , wherein the circuitry is further configured to demodulate the input 5G RF signal having the new RF impairment based on the trained neural network system.

10. The communication system according to claim 6 , wherein the circuitry is further configured to classify modulation associated with the input 5G RF signal based on the trained neural network system.

11. The communication system according to claim 10 , wherein the circuitry is further configured to detect an order of modulation of a physical downlink shared channel (PDSCH) and a physical uplink shared channel (PUSCH) from the input 5G RF signal based on the trained neural network system independent of decoding of data from the 5G RF signal.

12. The communication system according to claim 1 , wherein the circuitry is further configured to train the neural network system to perform channel equalization in presence of noise.

13. The communication system according to claim 1 , wherein the neural network system comprises a plurality of convolutional neural networks (CNN) layers followed by a Softmax layer.

14. The communication system according to claim 1 , wherein the neural network system comprises a combination of a plurality of convolutional neural networks (CNN) layers and a plurality of long short-term memory (LSTM) layers followed by Softmax layer.

15. The communication system according to claim 1 , wherein the neural network system comprises a combination of a plurality of convolutional neural networks (CNN) layers and a plurality of gated recurrent units (GRU) layers followed by Softmax layer.

16. A method, comprising:
in a communication system that comprises a neural network system and circuitry:
labelling, by the circuitry, a first set of fifth generation (5G) radio frequency (RF) signals originated from a first type of source and a second set of 5G RF signals originated from a second type of source;
artificially adding, by the circuitry, at least one RF impairment randomly to each labelled example of the first set of 5G RF signals and the second set of 5G RF signals, wherein randomization of artificially added RF impairment to each labeled example corresponds to addition of different RF impairments randomly to different labeled examples;
training, by the circuitry, the neural network system with a plurality of labelled examples, wherein each labelled example includes the artificially added RF impairment; and
executing a plurality of functionalities of a 5G modem based on the trained neural network system.

17. The method according to claim 16 , further comprising decoding, by the circuitry, synchronization signals from 5G new radio (NR) synchronization signal (SS) burst that includes a plurality of synchronization signal blocks (SSBs) based on the trained neural network system, wherein the 5G NR SS burst is transmitted by use of one or more directional beams by a base station and captured by the communication system to decode the synchronization signals based on the trained neural network system for time-frequency synchronization.

18. The method according to claim 16 , further comprising extracting physical cell identifier (ID) of a serving base station from an input 5G RF signal based on the trained neural network system.

19. The method according to claim 18 , further comprising:
demodulating, by the circuitry, the input 5G RF signal having a new RF impairment; and
classifying, by the circuitry, modulation associated with the input 5G RF signal.

20. The method according to claim 19 , further comprising demodulating, by the circuitry, an input 5G RF signal having the new RF impairment based on the trained neural network system.

Assignees

• Movandi Corp

Inventors

• Rofougaran Ahmadreza
• Gharavi Sam
• Liu Xianfu
• Zou Qiyue

Classifications

• H04W76/11

  Allocation or use of connection identifiers · CPC title

• H04B17/318

  Received signal strength · CPC title

• H04W24/10

  Scheduling measurement reports {; Arrangements for measurement reports} · CPC title

• H04W16/28

  using beam steering · CPC title

• H04W56/001 Primary

  Synchronization between nodes · CPC title