Systems and methods for wireless signal configuration by a neural network

Patent metadata

Field	Value
Publication number	US-2020366385-A1
Application number	US-201916412908-A
Country	US
Kind code	A1
Filing date	May 15, 2019
Priority date	May 15, 2019
Publication date	Nov 19, 2020
Grant date	—

Abstract

A wireless network can generate candidate signal configurations for physical transmissions to or from a user equipment (UE) in a radio environment. The generation of candidate signal configurations can be performed using a first neural network that is associated with the UE. These signal configurations can then be evaluated using a second neural network that is associated with the radio environment. The second neural network can be trained using measurements from previous physical transmissions in the radio environment. The trained second neural network generates a reward value that is associated with the candidate signal configurations. The first neural network is then trained using the reward values from the second neural network to produce improved candidate signal configurations. When a signal configuration that produces a suitable reward value is generated, this signal configuration can be used for the physical transmission in the radio environment.

First claim

1 . A method comprising:
collecting a plurality of data samples, each data sample comprising information associated with a respective physical transmission in a radio environment; training a first neural network associated with the radio environment using a subset of the plurality of data samples; receiving, from a second neural network associated with a user equipment (UE), a candidate signal configuration for a scheduled transmission in the radio environment; and evaluating, using the trained first neural network, the candidate signal configuration to produce an evaluation of the candidate signal configuration for training the second neural network.

2 . The method of claim 1 , wherein the UE is a first UE, and at least one data sample of the plurality of data samples comprises information associated with a physical transmission to or from a second UE in the radio environment.

3 . The method of claim 2 , wherein the information associated with the physical transmission to or from the second UE comprises:
a first state of the second UE before the physical transmission to or from the second UE; transmission parameters associated with the physical transmission to or from the second UE; a second state of the second UE after the physical transmission to or from the second UE; and a metric representing an effectiveness of the physical transmission to or from the second UE in the radio environment.

4 . The method of claim 3 , wherein the metric comprises a weighted sum of a plurality of measurements associated with the physical transmission to or from the second UE.

5 . The method of claim 1 , wherein collecting the plurality of data samples comprises:
receiving the plurality of data samples from a plurality of network devices; and storing the plurality of data samples in a database.

6 . The method of claim 5 , wherein storing the plurality of data samples in the database comprises deleting a plurality of older data samples from the database.

7 . The method of claim 5 , wherein the plurality of network devices comprises a base station and a plurality of UEs.

8 . The method of claim 1 , wherein the method further comprises:
receiving, from the second neural network, information associated with a state of the UE and information associated with the radio environment.

9 . The method of claim 8 , wherein:
evaluating the candidate signal configuration comprises inputting the candidate signal configuration, the information associated with the state of the UE and the information associated with the radio environment into the trained first neural network; and the evaluation of the candidate signal configuration comprises a metric representing a predicted effectiveness of the candidate signal configuration in the radio environment and information associated with a predicted state of the UE after performing the scheduled transmission to or from the UE using the candidate signal configuration.

10 . The method of claim 1 , wherein the UE is a first UE and the scheduled transmission is a first scheduled transmission, the method further comprising:
receiving, from a third neural network associated with a second UE, a candidate signal configuration for a second scheduled transmission in the radio environment; and evaluating, using the trained first neural network, the candidate signal configuration for the second scheduled transmission to produce an evaluation of the candidate signal configuration for the second scheduled transmission for training the third neural network.

11 . The method of claim 1 , wherein the method is performed at a base station.

12 . A method comprising:
generating, using a first neural network associated with a user equipment (UE), a candidate signal configuration for a scheduled transmission in a radio environment; receiving, from a second neural network associated with the radio environment, an evaluation of the candidate signal configuration; training the first neural network based on the evaluation of the candidate signal configuration; and generating, using the trained first neural network, a final signal configuration for the scheduled transmission in the radio environment.

13 . The method of claim 12 , further comprising:
sending the final signal configuration to a network device to perform the scheduled transmission in the radio environment.

14 . The method of claim 13 , wherein:
the network device comprises the UE and the scheduled transmission is performed from the UE; or the network device comprises a base station and the scheduled transmission is performed from the base station to the UE.

15 . The method of claim 13 , further comprising:
determining, after the network device performs the scheduled transmission, a plurality of measurements associated with the scheduled transmission; and sending the plurality of measurements to a database associated with the second neural network.

16 . The method of claim 12 , wherein training the first neural network comprises training the first neural network using a reinforcement learning model.

17 . The method of claim 12 , further comprising:
sending, to the second neural network, the candidate signal configuration, information associated with a state of the UE and information associated with the radio environment.

18 . The method of claim 12 , wherein the evaluation of the candidate signal configuration comprises a metric representing a predicted effectiveness of the candidate signal configuration in the radio environment and information associated with a predicted state of the UE after performing the scheduled transmission to or from the UE using the candidate signal configuration.

19 . The method of claim 12 , further comprising:
generating, using the trained first neural network, another final signal configuration for another scheduled transmission in the radio environment.

20 . The method of claim 12 , wherein the method is performed at a base station.

21 . A system comprising:
a processor; and at least one computer readable storage medium storing:
a database comprising a plurality of data samples, each data sample comprising information associated with a respective physical transmission in a radio environment;
a first neural network associated with the radio environment; and
programming for execution by the processor, the programming including instructions to perform actions in accordance with a method that comprises:
training the first neural network using a subset of the plurality of data samples;
receiving, from a second neural network associated with a user equipment (UE), candidate signal configuration for a scheduled transmission in the radio environment; and
evaluating, using the trained first neural network, the candidate signal configuration to produce an evaluation of the candidate signal configuration for training the second neural network.

22 . A system comprising:
a processor; and at least one computer readable storage medium storing:
a first neural network associated with a user equipment (UE); and
programming for execution by the processor, the programming including instructions to perform actions in accordance with a method that comprises:
generating, using the first neural network, a candidate signal configuration for a scheduled transmission in a radio environment;
receiving, from a second neural network associated with the radio environment, an evaluation of the candidate signal configuration;
training the firs

Assignees

• Ge Yiqun
• Shi Wuxian
• Tong Wen
• Huawei Tech Co Ltd

Inventors

• Ge Yiqun
• Shi Wuxian
• Tong Wen

Classifications

• G06N3/048

  Activation functions · CPC title

• G06N3/045

  Combinations of networks · CPC title

• G06N3/044

  Recurrent networks, e.g. Hopfield networks · CPC title

• G06N3/09

  Supervised learning · CPC title

• G06N3/0442

  characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU] · CPC title