Granted Patent B2

US 12,554,325 · App. 17/741,263 · Granted Feb 17, 2026

Inventors: Nadine Hussami (Brooklyn, NY); Patrick William John Kaifosh (New York, NY); Alexandre Barachant (Brooklyn, NY); Daniel Wetmore (Brooklyn, NY)

Assignee: Meta Platforms Technologies, LLC

G06F3/015G06F3/017G06F3/02G06F3/0487

Abstract

Claims (39)

1 . A computer-implemented method comprising:

accessing neuromuscular sensor data at a first time generated by a neuromuscular sensor on a wearable device donned by a user;

inputting the neuromuscular sensor data accessed at the first time into a trained inferential model that is configured to predict body state of a body part of the user;

predicting, based on the neuromuscular sensor data accessed at the first time, body state information for a second time that is a specified time interval after the first time for the body part of the user based on one or more outputs of the trained inferential model such that a temporal latency between the predicted body state information for the second time and an actual body state for the second time is less than a threshold amount of latency;

generating a visual representation of the body part of the user based on the predicted body state information; and

displaying the visual representation of the body part of the user via a pair of augmented-reality glasses.

2 . The computer-implemented method of claim 1 , wherein the wearable device includes a plurality of neuromuscular sensors arranged in a circumferential array.

3 . The computer-implemented method of claim 2 , wherein the plurality of neuromuscular sensors of the wearable device record neuromuscular signals from the user as the user exerts force or performs at least one of movements, poses, or gestures.

4 . The computer-implemented method of claim 1 , wherein the wearable device includes one or more auxiliary sensors configured to continuously record auxiliary signals that are implemented as inputs to the trained inferential model.

5 . The computer-implemented method of claim 1 , wherein the body state information is further predicted based on derived signal data.

6 . The computer-implemented method of claim 5 , wherein the derived signal data is integrated or filtered to determine movement of one or more muscles of the user during performance of a gesture.

7 . The computer-implemented method of claim 1 , wherein the neuromuscular sensor data generated by the neuromuscular sensor on the wearable device represents a discrete gesture performed by the user.

8 . The computer-implemented method of claim 7 , wherein the visual representation of the body part of the user includes the discrete gesture.

9 . The computer-implemented method of claim 1 , wherein the neuromuscular sensor data generated by the neuromuscular sensor on the wearable device represents a continuous movement gesture performed by the user.

10 . The computer-implemented method of claim 9 , wherein inputting the neuromuscular sensor data into the trained inferential model includes providing continuous, real-time inputs to the inferential model as neuromuscular signals generated by the user are being recorded.

11 . The computer-implemented method of claim 9 , wherein body state information is predicted for the user's body part in real-time, resulting in a real-time estimation of positions or forces of the user's body part.

12 . The computer-implemented method of claim 11 , wherein the visual representation of the users' body part includes the continuous movement gesture.

13 . A system comprising:

at least one physical processor; and

physical memory comprising computer-executable instructions that, when executed by the physical processor, cause the physical processor to:

access neuromuscular sensor data at a first time generated by a neuromuscular sensor on a wearable device donned by a user;

input the neuromuscular sensor data accessed at the first time into a trained inferential model that is configured to predict body state of a body part of the user;

predict, based on the neuromuscular sensor data accessed at the first time, body state information for a second time that is a specified time interval after the first time for the body part of the user based on one or more outputs of the trained inferential model such that a temporal latency between the predicted body state information for the second time and an actual body state for the second time is less than a threshold amount of latency;

generate a visual representation of the body part of the user based on the predicted body state information; and

display the visual representation of the body part of the user via a pair of augmented-reality glasses.

14 . The system of claim 13 , wherein the wearable device includes a plurality of neuromuscular sensors arranged in a circumferential array, and the plurality of neuromuscular sensors of the wearable device record neuromuscular signals from the user as the user exerts force or performs at least one of movements, poses, or gestures.

15 . The system of claim 13 , wherein the wearable device includes one or more auxiliary sensors configured to continuously record auxiliary signals that are implemented as inputs to the trained inferential model.

16 . The system of claim 13 , wherein the body state information is further predicted based on derived signal data.

17 . The system of claim 16 , wherein the derived signal data is integrated or filtered to determine movement of one or more muscles during performance of a gesture.

18 . The system of claim 13 , wherein the neuromuscular sensor data generated by the neuromuscular sensor on the wearable device represents a discrete gesture performed by the user.

19 . A non-transitory computer-readable medium comprising one or more computer-executable instructions that, when executed by at least one processor of a computing device, cause the computing device to:

access neuromuscular sensor data at a first time generated by a neuromuscular sensor on a wearable device donned by a user;

input the neuromuscular sensor data accessed at the first time into a trained inferential model that is configured to predict body state of a body part of the user;

predict, based on the neuromuscular sensor data accessed at the first time, body state information for a second time that is a specified time interval after the first time for the body part of the user based on one or more outputs of the trained inferential model such that a temporal latency between the predicted body state information for the second time and an actual body state for the second time is less than a threshold amount of latency;

generate a visual representation of the body part of the user based on the predicted body state information; and

display the visual representation of the body part of the user via a pair of augmented-reality glasses.

20 . The method of claim 1 , further comprising:

accessing inertial sensor data generated by an inertial measuring unit, wherein the inertial sensor data is distinct from the neuromuscular sensor data and the inertial measuring unit is distinct from the neuromuscular sensor; and

wherein inputting the neuromuscular sensor data accessed at the first time into the trained inferential model includes inputting the inertial sensor data into the trained inferential model.

Assignments (4)

Continuity (45)

References Cited (400)