Deep Reinforcement Learning for Control of Time-Varying Musculoskeletal Systems With High Fatigability: A Feasibility Study

Functional electrical stimulation (FES) can be used to restore motor function to people with paralysis caused by spinal cord injuries (SCIs). However, chronically-paralyzed FES-stimulated muscles can fatigue quickly, which may decrease FES controller performance. In this work, we explored the feasib...

Full description

Saved in:
Bibliographic Details
Main Authors: Jessica Abreu, Douglas C. Crowder, Robert F. Kirsch
Format: Article
Language:English
Published: IEEE 2022-01-01
Series:IEEE Transactions on Neural Systems and Rehabilitation Engineering
Subjects:
Online Access:https://ieeexplore.ieee.org/document/9875337/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849763199043239936
author Jessica Abreu
Douglas C. Crowder
Robert F. Kirsch
author_facet Jessica Abreu
Douglas C. Crowder
Robert F. Kirsch
author_sort Jessica Abreu
collection DOAJ
description Functional electrical stimulation (FES) can be used to restore motor function to people with paralysis caused by spinal cord injuries (SCIs). However, chronically-paralyzed FES-stimulated muscles can fatigue quickly, which may decrease FES controller performance. In this work, we explored the feasibility of using deep neural network (DNN) controllers trained with reinforcement learning (RL) to control FES of upper-limb muscles after SCI. We developed upper-limb biomechanical models that exhibited increased muscle fatigability, decreased muscle recovery, and decreased muscle strength, as observed in people with chronic SCIs. Simulations confirmed that controller training time and controller performance are impaired to varying degrees by muscle fatigability. Also, the simulations showed that large muscle strength asymmetries between opposing muscles can substantially impair controller performance. However, the results of this study suggest that controller performance for highly-fatigable musculoskeletal systems can be preserved by allowing for rest between movements. Overall, the results suggest that RL can be used to successfully train FES controllers, even for highly-fatigable musculoskeletal systems.
format Article
id doaj-art-7960bcae7a1d4509a0fdf330e43a1d7a
institution DOAJ
issn 1534-4320
1558-0210
language English
publishDate 2022-01-01
publisher IEEE
record_format Article
series IEEE Transactions on Neural Systems and Rehabilitation Engineering
spelling doaj-art-7960bcae7a1d4509a0fdf330e43a1d7a2025-08-20T03:05:29ZengIEEEIEEE Transactions on Neural Systems and Rehabilitation Engineering1534-43201558-02102022-01-01302613262210.1109/TNSRE.2022.32039709875337Deep Reinforcement Learning for Control of Time-Varying Musculoskeletal Systems With High Fatigability: A Feasibility StudyJessica Abreu0https://orcid.org/0000-0002-5743-4489Douglas C. Crowder1https://orcid.org/0000-0002-4435-1799Robert F. Kirsch2https://orcid.org/0000-0003-2564-1800Department of Biomedical Engineering, Case Western Reserve University, Cleveland, OH, USACognitive and Emerging Computing Department, Sandia National Laboratories, Albuquerque, NM, USADepartment of Biomedical Engineering, Case Western Reserve University, Cleveland, OH, USAFunctional electrical stimulation (FES) can be used to restore motor function to people with paralysis caused by spinal cord injuries (SCIs). However, chronically-paralyzed FES-stimulated muscles can fatigue quickly, which may decrease FES controller performance. In this work, we explored the feasibility of using deep neural network (DNN) controllers trained with reinforcement learning (RL) to control FES of upper-limb muscles after SCI. We developed upper-limb biomechanical models that exhibited increased muscle fatigability, decreased muscle recovery, and decreased muscle strength, as observed in people with chronic SCIs. Simulations confirmed that controller training time and controller performance are impaired to varying degrees by muscle fatigability. Also, the simulations showed that large muscle strength asymmetries between opposing muscles can substantially impair controller performance. However, the results of this study suggest that controller performance for highly-fatigable musculoskeletal systems can be preserved by allowing for rest between movements. Overall, the results suggest that RL can be used to successfully train FES controllers, even for highly-fatigable musculoskeletal systems.https://ieeexplore.ieee.org/document/9875337/Reinforcement learningfunctional electrical stimulationmotor controlspinal cord injurybiomechanical model
spellingShingle Jessica Abreu
Douglas C. Crowder
Robert F. Kirsch
Deep Reinforcement Learning for Control of Time-Varying Musculoskeletal Systems With High Fatigability: A Feasibility Study
IEEE Transactions on Neural Systems and Rehabilitation Engineering
Reinforcement learning
functional electrical stimulation
motor control
spinal cord injury
biomechanical model
title Deep Reinforcement Learning for Control of Time-Varying Musculoskeletal Systems With High Fatigability: A Feasibility Study
title_full Deep Reinforcement Learning for Control of Time-Varying Musculoskeletal Systems With High Fatigability: A Feasibility Study
title_fullStr Deep Reinforcement Learning for Control of Time-Varying Musculoskeletal Systems With High Fatigability: A Feasibility Study
title_full_unstemmed Deep Reinforcement Learning for Control of Time-Varying Musculoskeletal Systems With High Fatigability: A Feasibility Study
title_short Deep Reinforcement Learning for Control of Time-Varying Musculoskeletal Systems With High Fatigability: A Feasibility Study
title_sort deep reinforcement learning for control of time varying musculoskeletal systems with high fatigability a feasibility study
topic Reinforcement learning
functional electrical stimulation
motor control
spinal cord injury
biomechanical model
url https://ieeexplore.ieee.org/document/9875337/
work_keys_str_mv AT jessicaabreu deepreinforcementlearningforcontroloftimevaryingmusculoskeletalsystemswithhighfatigabilityafeasibilitystudy
AT douglasccrowder deepreinforcementlearningforcontroloftimevaryingmusculoskeletalsystemswithhighfatigabilityafeasibilitystudy
AT robertfkirsch deepreinforcementlearningforcontroloftimevaryingmusculoskeletalsystemswithhighfatigabilityafeasibilitystudy