Unlocking Gait Analysis Beyond the Gait Lab: High-Fidelity Replication of Knee Kinematics Using Inertial Motion Units and a Convolutional Neural Network

Background: Gait analysis using three-dimensional motion capture systems (3D motion capture) provides a combination of kinematic and kinetic measurements for quantifying and characterizing the motion and loads, respectively, of lower extremity joints during human movement. However, their high cost a...

Full description

Saved in:
Bibliographic Details
Main Authors: Stefano A. Bini, MD, Nicholas Gillian, PhD, Thomas A. Peterson, PhD, Richard B. Souza, PhD, PT, Brooke Schultz, MS, ACE-CPT, Wojciech Mormul, MS, Marek K. Cichoń, MS, Agnieszka Barbara Szczotka, MS, Ivan Poupyrev, PhD
Format: Article
Language:English
Published: Elsevier 2025-06-01
Series:Arthroplasty Today
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S2352344125000433
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849434709734457344
author Stefano A. Bini, MD
Nicholas Gillian, PhD
Thomas A. Peterson, PhD
Richard B. Souza, PhD, PT
Brooke Schultz, MS, ACE-CPT
Wojciech Mormul, MS
Marek K. Cichoń, MS
Agnieszka Barbara Szczotka, MS
Ivan Poupyrev, PhD
author_facet Stefano A. Bini, MD
Nicholas Gillian, PhD
Thomas A. Peterson, PhD
Richard B. Souza, PhD, PT
Brooke Schultz, MS, ACE-CPT
Wojciech Mormul, MS
Marek K. Cichoń, MS
Agnieszka Barbara Szczotka, MS
Ivan Poupyrev, PhD
author_sort Stefano A. Bini, MD
collection DOAJ
description Background: Gait analysis using three-dimensional motion capture systems (3D motion capture) provides a combination of kinematic and kinetic measurements for quantifying and characterizing the motion and loads, respectively, of lower extremity joints during human movement. However, their high cost and limited accessibility impact their utility. Wearable inertial motion sensors offer a cost-effective alternative to measure simple temporospatial variables, but more complex kinematic variables require machine learning interfaces. We hypothesize that kinematic measures about the knee collected using motion capture can be replicated by coupling raw data collected from inertial measurement units (IMUs) to machine learning algorithms. Methods: Data from 40 healthy participants performing fixed walking, stair climbing, and sit-to-stand tasks were collected using both 3D motion capture and IMUs. Sequence to sequence convolutional neural networks were trained to map IMU data to three motion capture kinematic outputs: right knee angle, right knee angular velocity, and right hip angle. Model performance was assessed using mean absolute error. Results: The convolutional neural network models exhibited high accuracy in replicating motion capture-derived kinematic variables. Mean absolute error values for right knee angle ranged from 4.30 ± 1.55 to 5.79 ± 2.93 degrees, for right knee angular velocity from 7.82 ± 3.01 to 22.16 ± 9.52 degrees per second, and for right hip angle from 4.82 ± 2.29 to 8.63 ± 4.73 degrees. Task-specific variations in accuracy were observed. Conclusions: The findings highlight the potential of leveraging raw data from wearable inertial sensors and machine learning algorithms to reproduce gait lab-quality kinematic data outside the laboratory settings for the study of knee function following joint injury, surgery, or the progression of joint disease.
format Article
id doaj-art-d468d774ba924dca99fbbb2c754ad07e
institution Kabale University
issn 2352-3441
language English
publishDate 2025-06-01
publisher Elsevier
record_format Article
series Arthroplasty Today
spelling doaj-art-d468d774ba924dca99fbbb2c754ad07e2025-08-20T03:26:33ZengElsevierArthroplasty Today2352-34412025-06-013310165610.1016/j.artd.2025.101656Unlocking Gait Analysis Beyond the Gait Lab: High-Fidelity Replication of Knee Kinematics Using Inertial Motion Units and a Convolutional Neural NetworkStefano A. Bini, MD0Nicholas Gillian, PhD1Thomas A. Peterson, PhD2Richard B. Souza, PhD, PT3Brooke Schultz, MS, ACE-CPT4Wojciech Mormul, MS5Marek K. Cichoń, MS6Agnieszka Barbara Szczotka, MS7Ivan Poupyrev, PhD8Department of Orthopaedic Surgery, University of California San Francisco, San Francisco, CA, USA; Corresponding author. Department of Orthopaedics, University of California San Francisco, 500 Parnassus Avenue, MU 323-W, San Francisco, CA 94143.Google Advanced Technology & Projects (ATAP) Invention Studio, Palo Alto, CA, USADepartment of Orthopaedic Surgery, University of California San Francisco, San Francisco, CA, USA; Bakar Computational Health Sciences Institute, University of California San Francisco, San Francisco, CA, USADepartment of Orthopaedic Surgery, University of California San Francisco, San Francisco, CA, USA; Department of Physical Therapy and Rehabilitation Science, University of California San Francisco, San Francisco, CA, USADepartment of Orthopaedic Surgery, University of California San Francisco, San Francisco, CA, USADepartment of Physical Therapy and Rehabilitation Science, University of California San Francisco, San Francisco, CA, USADepartment of Physical Therapy and Rehabilitation Science, University of California San Francisco, San Francisco, CA, USADepartment of Physical Therapy and Rehabilitation Science, University of California San Francisco, San Francisco, CA, USADepartment of Physical Therapy and Rehabilitation Science, University of California San Francisco, San Francisco, CA, USABackground: Gait analysis using three-dimensional motion capture systems (3D motion capture) provides a combination of kinematic and kinetic measurements for quantifying and characterizing the motion and loads, respectively, of lower extremity joints during human movement. However, their high cost and limited accessibility impact their utility. Wearable inertial motion sensors offer a cost-effective alternative to measure simple temporospatial variables, but more complex kinematic variables require machine learning interfaces. We hypothesize that kinematic measures about the knee collected using motion capture can be replicated by coupling raw data collected from inertial measurement units (IMUs) to machine learning algorithms. Methods: Data from 40 healthy participants performing fixed walking, stair climbing, and sit-to-stand tasks were collected using both 3D motion capture and IMUs. Sequence to sequence convolutional neural networks were trained to map IMU data to three motion capture kinematic outputs: right knee angle, right knee angular velocity, and right hip angle. Model performance was assessed using mean absolute error. Results: The convolutional neural network models exhibited high accuracy in replicating motion capture-derived kinematic variables. Mean absolute error values for right knee angle ranged from 4.30 ± 1.55 to 5.79 ± 2.93 degrees, for right knee angular velocity from 7.82 ± 3.01 to 22.16 ± 9.52 degrees per second, and for right hip angle from 4.82 ± 2.29 to 8.63 ± 4.73 degrees. Task-specific variations in accuracy were observed. Conclusions: The findings highlight the potential of leveraging raw data from wearable inertial sensors and machine learning algorithms to reproduce gait lab-quality kinematic data outside the laboratory settings for the study of knee function following joint injury, surgery, or the progression of joint disease.http://www.sciencedirect.com/science/article/pii/S2352344125000433Knee kinematicsInertial motion sensorsMachine learningGait analysisMusculoskeletal researchWearables
spellingShingle Stefano A. Bini, MD
Nicholas Gillian, PhD
Thomas A. Peterson, PhD
Richard B. Souza, PhD, PT
Brooke Schultz, MS, ACE-CPT
Wojciech Mormul, MS
Marek K. Cichoń, MS
Agnieszka Barbara Szczotka, MS
Ivan Poupyrev, PhD
Unlocking Gait Analysis Beyond the Gait Lab: High-Fidelity Replication of Knee Kinematics Using Inertial Motion Units and a Convolutional Neural Network
Arthroplasty Today
Knee kinematics
Inertial motion sensors
Machine learning
Gait analysis
Musculoskeletal research
Wearables
title Unlocking Gait Analysis Beyond the Gait Lab: High-Fidelity Replication of Knee Kinematics Using Inertial Motion Units and a Convolutional Neural Network
title_full Unlocking Gait Analysis Beyond the Gait Lab: High-Fidelity Replication of Knee Kinematics Using Inertial Motion Units and a Convolutional Neural Network
title_fullStr Unlocking Gait Analysis Beyond the Gait Lab: High-Fidelity Replication of Knee Kinematics Using Inertial Motion Units and a Convolutional Neural Network
title_full_unstemmed Unlocking Gait Analysis Beyond the Gait Lab: High-Fidelity Replication of Knee Kinematics Using Inertial Motion Units and a Convolutional Neural Network
title_short Unlocking Gait Analysis Beyond the Gait Lab: High-Fidelity Replication of Knee Kinematics Using Inertial Motion Units and a Convolutional Neural Network
title_sort unlocking gait analysis beyond the gait lab high fidelity replication of knee kinematics using inertial motion units and a convolutional neural network
topic Knee kinematics
Inertial motion sensors
Machine learning
Gait analysis
Musculoskeletal research
Wearables
url http://www.sciencedirect.com/science/article/pii/S2352344125000433
work_keys_str_mv AT stefanoabinimd unlockinggaitanalysisbeyondthegaitlabhighfidelityreplicationofkneekinematicsusinginertialmotionunitsandaconvolutionalneuralnetwork
AT nicholasgillianphd unlockinggaitanalysisbeyondthegaitlabhighfidelityreplicationofkneekinematicsusinginertialmotionunitsandaconvolutionalneuralnetwork
AT thomasapetersonphd unlockinggaitanalysisbeyondthegaitlabhighfidelityreplicationofkneekinematicsusinginertialmotionunitsandaconvolutionalneuralnetwork
AT richardbsouzaphdpt unlockinggaitanalysisbeyondthegaitlabhighfidelityreplicationofkneekinematicsusinginertialmotionunitsandaconvolutionalneuralnetwork
AT brookeschultzmsacecpt unlockinggaitanalysisbeyondthegaitlabhighfidelityreplicationofkneekinematicsusinginertialmotionunitsandaconvolutionalneuralnetwork
AT wojciechmormulms unlockinggaitanalysisbeyondthegaitlabhighfidelityreplicationofkneekinematicsusinginertialmotionunitsandaconvolutionalneuralnetwork
AT marekkcichonms unlockinggaitanalysisbeyondthegaitlabhighfidelityreplicationofkneekinematicsusinginertialmotionunitsandaconvolutionalneuralnetwork
AT agnieszkabarbaraszczotkams unlockinggaitanalysisbeyondthegaitlabhighfidelityreplicationofkneekinematicsusinginertialmotionunitsandaconvolutionalneuralnetwork
AT ivanpoupyrevphd unlockinggaitanalysisbeyondthegaitlabhighfidelityreplicationofkneekinematicsusinginertialmotionunitsandaconvolutionalneuralnetwork