Granted Patent B2

US 12,652,602 · App. 17/305,119 · Granted Jun 9, 2026

Inventors: Anantharaman Balasubramanian (San Diego, CA); Shuanshuan Wu (San Diego, CA); Kapil Gulati (Belle Mead, NJ); Kyle Chi Guan (New York, NY); Himaja Kesavareddigari (Bridgewater, NJ)

Assignee: QUALCOMM Incorporated

H04W40/18G06N5/04G06N20/00H04B17/3913H04L1/0001

Abstract

Claims (94)

1 . A method of wireless communication performed by a transmitter node, comprising:

receiving, from an assisting wireless node, information configuring a machine learning model for predicting a future state associated with a wireless channel at a future time instance, wherein the information configuring the machine learning model comprises one or more weights associated with the machine learning model and one or more architectural parameters for the machine learning model;

receiving, from the assisting wireless node, one or more parameters for detecting a machine learning outage event, wherein the one or more parameters for detecting the machine learning outage event include a mean channel prediction error over a time period or a number of transmissions that resulted in incorrect hybrid automatic repeat request feedback;

transmitting, to the assisting wireless node, a request to update the machine learning model based at least in part on detecting the machine learning outage event in accordance with the one or more parameters for detecting the machine learning outage event;

receiving, from the assisting wireless node, one or more updates to the one or more architectural parameters associated with the machine learning model;

predicting the future state using the machine learning model associated with the one or more updates to the one or more architectural parameters associated with the machine learning model, and based at least in part on one or more inputs to the machine learning model, wherein the one or more inputs comprise a current state associated with the wireless channel or one or more previous states associated with the wireless channel;

selecting one or more parameters for a transmission to occur at the future time instance based at least in part on the future state associated with the wireless channel; and

transmitting the transmission using the one or more parameters.

2 . The method of claim 1 ,

wherein the model update comprises a periodic model update.

3 . The method of claim 1 ,

wherein the current state or the one or more previous states associated with the wireless channel include one or more of a measurement associated with a received wireless signal, a number of sensed devices in a vicinity of the transmitter node, or a number of the sensed devices in the vicinity of the transmitter node that are moving at a speed that satisfies a threshold.

4 . The method of claim 1 ,

wherein predicting the future state associated with the wireless channel includes:

providing, to the machine learning model at one or more layers, one or more inputs that relate to the current state associated with the wireless channel or the one or more previous states associated with the wireless channel; and

obtaining, from the machine learning model, an output that indicates the future state associated with the wireless channel based at least in part on the weights and an activation function associated with the machine learning model at the one or more layers.

5 . The method of claim 1 ,

wherein the one or more parameters for the transmission include one or more of a number of retransmissions of the transmission, a modulation order, a code rate, a number of transmit antennas, a precoder, or a transmit power.

6 . The method of claim 1 , further comprising:

receiving, from an assisting wireless node, information configuring the machine learning model based at least in part on the wireless channel having a semi-static configuration in which devices have fixed relative positions or predictable motion patterns.

7 . The method of claim 1 ,

wherein the one or more updates to one or more architectural parameters comprises one or more of: one or more updates to a number of layers associated with the machine learning model, one or more updates to one or more activation functions used in different layers of the machine learning model, one or more updates to one or more weights associated with different layers of the machine learning model, one or more updates to one or more drop out parameters associated with the machine learning model, one or more updates to one or more architectural connection parameters for one or more cell gates or forget gates associated with the machine learning model, or one or more updates to one or more inputs to an inference algorithm associated with the machine learning model.

8 . The method of claim 1 , wherein the wireless channel is between the transmitter node and a receiver node.

9 . The method of claim 8 , further comprising:

transmitting the transmission to the receiver node over the wireless channel.

10 . The method of claim 8 , wherein the wireless channel is a sidelink channel.

11 . The method of claim 1 , further comprising:

receiving, from the assisting wireless node, in accordance with the request, an update to the one or more parameters for detecting the machine learning outage event.

12 . A transmitter node for wireless communication, comprising:

one or more memories; and

one or more processors, coupled to the one or more memories, configured to:

receive, from an assisting wireless node, information to configure a machine learning model for predicting a future state associated with a wireless channel at a future time instance, wherein the information configuring the machine learning model comprises one or more weights associated with the machine learning model and one or more architectural parameters for the machine learning model;

receive, from the assisting wireless node, one or more parameters for detecting a machine learning outage event, wherein the one or more parameters for detecting the machine learning outage event include a mean channel prediction error over a time period or a number of transmissions that resulted in incorrect hybrid automatic repeat request feedback;

transmit, to the assisting wireless node, a request to update the machine learning model based at least in part on detecting the machine learning outage event in accordance with the one or more parameters for detecting the machine learning outage event;

receive, from the assisting wireless node, one or more updates to the one or more architectural parameters associated with the machine learning model;

predict the future state using the machine learning model associated with the one or more updates to the one or more architectural parameters associated with the machine learning model, and based at least in part on one or more inputs to the machine learning model, wherein the one or more inputs comprise a current state associated with the wireless channel or one or more previous states associated with the wireless channel;

select one or more parameters for a transmission to occur at the future time instance based at least in part on the future state associated with the wireless channel; and

transmit the transmission using the one or more parameters.

13 . The transmitter node of claim 12 ,

wherein the model update comprises a periodic model update.

14 . The transmitter node of claim 12 ,

wherein the current state or the one or more previous states associated with the wireless channel include one or more of a measurement associated with a received wireless signal, a number of sensed devices in a vicinity of the transmitter node, or a number of the sensed devices in the vicinity of the transmitter node that are moving at a speed that satisfies a threshold.

15 . The transmitter node of claim 12 ,

wherein the one or more processors, to predict the future state associated with the wireless channel, are configured to:

provide, to the machine learning model at one or more layers, one or more inputs that relate to the current state associated with the wireless channel or the one or more previous states associated with the wireless channel; and

obtain, from the machine learning model, an output that indicates the future state associated with the wireless channel based at least in part on the weights and an activation function associated with the machine learning model at the one or more layers.

16 . The transmitter node of claim 12 ,

wherein the one or more parameters for the transmission include one or more of a number of retransmissions of the transmission, a modulation order, a code rate, a number of transmit antennas, a precoder, or a transmit power.

17 . The transmitter node of claim 12 ,

wherein the one or more processors are further configured to:

receive, from an assisting wireless node, information configuring the machine learning model based at least in part on the wireless channel having a semi-static configuration in which devices have fixed relative positions or predictable motion patterns.

18 . The transmitter node of claim 12 ,

wherein the one or more updates to one or more architectural parameters comprises one or more of: one or more updates to a number of layers associated with the machine learning model, one or more updates to one or more activation functions used in different layers of the machine learning model, one or more updates to one or more weights associated with different layers of the machine learning model, one or more updates to one or more drop out parameters associated with the machine learning model, one or more updates to one or more architectural connection parameters for one or more cell gates or forget gates associated with the machine learning model, or one or more updates to one or more inputs to an inference algorithm associated with the machine learning model.

19 . The transmitter node of claim 12 , wherein the one or more processors are further configured to:

receive, from the assisting wireless node, in accordance with the request, an update to the one or more parameters for detecting the machine learning outage event.

20 . A non-transitory computer-readable medium storing a set of instructions for wireless communication, the set of instructions comprising:

one or more instructions that, when executed by one or more processors of a transmitter node, cause the transmitter node to:

receive, from an assisting wireless node, information to configure a machine learning model for predicting a future state associated with a wireless channel at a future time instance, wherein the information configuring the machine learning model comprises one or more weights associated with the machine learning model and one or more architectural parameters for the machine learning model;

receive, from the assisting wireless node, one or more parameters for detecting a machine learning outage event, wherein the one or more parameters for detecting the machine learning outage event include a mean channel prediction error over a time period or a number of transmissions that resulted in incorrect hybrid automatic repeat request feedback;

transmit, to the assisting wireless node, a request to update the machine learning model based at least in part on detecting the machine learning outage event in accordance with the one or more parameters for detecting the machine learning outage event;

receive, from the assisting wireless node, one or more updates to the one or more architectural parameters associated with the machine learning model;

predict the future state using the machine learning model associated with the one or more updates to the one or more architectural parameters associated with the machine learning model, and based at least in part on one or more inputs to the machine learning model, wherein the one or more inputs comprise a current state associated with the wireless channel or one or more previous states associated with the wireless channel;

select one or more parameters for a transmission to occur at the future time instance based at least in part on the future state associated with the wireless channel; and

transmit the transmission using the one or more parameters.

21 . The non-transitory computer-readable medium of claim 20 ,

wherein the model update comprises a periodic model update.

22 . The non-transitory computer-readable medium of claim 20 ,

wherein the current state or the one or more previous states associated with the wireless channel include one or more of a measurement associated with a received wireless signal, a number of sensed devices in a vicinity of the transmitter node, or a number of the sensed devices in the vicinity of the transmitter node that are moving at a speed that satisfies a threshold.

23 . The non-transitory computer-readable medium of claim 20 ,

wherein the one or more instructions, to predict the future state associated with the wireless channel, are further executable to cause the transmitter node to:

provide, to the machine learning model at one or more layers, one or more inputs that relate to the current state associated with the wireless channel or the one or more previous states associated with the wireless channel; and

obtain, from the machine learning model, an output that indicates the future state associated with the wireless channel based at least in part on the weights and an activation function associated with the machine learning model at the one or more layers.

24 . The non-transitory computer-readable medium of claim 20 ,

wherein the one or more parameters for the transmission include one or more of a number of retransmissions of the transmission, a modulation order, a code rate, a number of transmit antennas, a precoder, or a transmit power.

25 . The non-transitory computer-readable medium of claim 20 ,

wherein the one or more instructions are further executable to cause the transmitter node to:

receive, from an assisting wireless node, information configuring the machine learning model based at least in part on the wireless channel having a semi-static configuration in which devices have fixed relative positions or predictable motion patterns.

26 . The non-transitory computer-readable medium of claim 20 ,

wherein the one or more updates to one or more architectural parameters comprises one or more of: one or more updates to a number of layers associated with the machine learning model, one or more updates to one or more activation functions used in different layers of the machine learning model, one or more updates to one or more weights associated with different layers of the machine learning model, one or more updates to one or more drop out parameters associated with the machine learning model, one or more updates to one or more architectural connection parameters for one or more cell gates or forget gates associated with the machine learning model, or one or more updates to one or more inputs to an inference algorithm associated with the machine learning model.

27 . An apparatus for wireless communication, comprising:

means for receiving, from an assisting wireless node, information configuring a machine learning model for predicting a future state associated with a wireless channel at a future time instance, wherein the information configuring the machine learning model comprises one or more weights associated with the machine learning model and one or more architectural parameters for the machine learning model;

means for receiving, from the assisting wireless node, one or more parameters for detecting a machine learning outage event, wherein the one or more parameters for detecting the machine learning outage event include a mean channel prediction error over a time period or a number of transmissions that resulted in incorrect hybrid automatic repeat request feedback;

means for transmitting, to the assisting wireless node, a request to update the machine learning model based at least in part on detecting the machine learning outage event in accordance with the one or more parameters for detecting the machine learning outage event;

means for receiving, from the assisting wireless node, one or more updates to the one or more architectural parameters associated with the machine learning model;

means for predicting the future state using the machine learning model associated with the one or more updates to the one or more architectural parameters associated with the machine learning model, and based at least in part on one or more inputs to the machine learning model, wherein the one or more inputs comprise a current state associated with the wireless channel or one or more previous states associated with the wireless channel;

means for selecting one or more parameters for a transmission to occur at the future time instance based at least in part on the future state associated with the wireless channel; and

means for transmitting the transmission using the one or more parameters.

28 . The apparatus of claim 27 ,

wherein the model update comprises a periodic model update.

29 . The apparatus of claim 27 ,

wherein the current state or the one or more previous states associated with the wireless channel include one or more of a measurement associated with a received wireless signal, a number of sensed devices in a vicinity of the apparatus, or a number of the sensed devices in the vicinity of the apparatus that are moving at a speed that satisfies a threshold.

30 . The non-transitory computer-readable medium of claim 20 ,

wherein the one or more instructions are further executable to cause the transmitter node to:

receive, from the assisting wireless node, in accordance with the request, an update to the one or more parameters for detecting the machine learning outage event.

Assignments (1)

Continuity (1)

References Cited (34)