US 12,277,272 B2
Systems and methods for electromyogram-based control using neural networks
Ronald James Cotton, Chicago, IL (US)
Assigned to Rehabilitation Institute of Chicago, Chicago, IL (US)
Appl. No. 17/769,394
Filed by REHABILITATION INSTITUTE OF CHICAGO, Chicago, IL (US)
PCT Filed Oct. 16, 2020, PCT No. PCT/US2020/056143
§ 371(c)(1), (2) Date Apr. 15, 2022,
PCT Pub. No. WO2021/077009, PCT Pub. Date Apr. 22, 2021.
Claims priority of provisional application 62/916,129, filed on Oct. 16, 2019.
Prior Publication US 2024/0184362 A1, Jun. 6, 2024
Int. Cl. G06F 3/01 (2006.01); A61F 4/00 (2006.01); G06N 3/084 (2023.01); G06N 3/09 (2023.01)
CPC G06F 3/015 (2013.01) [A61F 4/00 (2013.01); G06N 3/084 (2013.01); G06N 3/09 (2023.01)] 15 Claims
OG exemplary drawing
 
1. A system for electromyography (EMG)-based control using neural networks, comprising:
one or more wearable sensors that generate EMG data; and
a computing device implementing a decoder including a neural network trained to decode EMG data generated in response to muscle activity of a user to one or more control signals configured to continuously control a predetermined device;
wherein the neural network is trained using a training set comprising a target location of each of a plurality of calibration targets and training EMG data generated in response to the user performing one or more self-selected movements indicative of movement to the target location of each of the plurality of calibration targets.