Recurrent Neural Network Methods for Extracting Dynamic Balance Variables during Gait from a Single Inertial Measurement Unit
Monitoring dynamic balance during gait is critical for fall prevention in the elderly. The current study aimed to develop recurrent neural network models for extracting balance variables from a single inertial measurement unit (IMU) placed on the sacrum during walking. Thirteen healthy young and thi...
Main Authors: | , , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2023-11-01
|
Series: | Sensors |
Subjects: | |
Online Access: | https://www.mdpi.com/1424-8220/23/22/9040 |
_version_ | 1797457850547568640 |
---|---|
author | Cheng-Hao Yu Chih-Ching Yeh Yi-Fu Lu Yi-Ling Lu Ting-Ming Wang Frank Yeong-Sung Lin Tung-Wu Lu |
author_facet | Cheng-Hao Yu Chih-Ching Yeh Yi-Fu Lu Yi-Ling Lu Ting-Ming Wang Frank Yeong-Sung Lin Tung-Wu Lu |
author_sort | Cheng-Hao Yu |
collection | DOAJ |
description | Monitoring dynamic balance during gait is critical for fall prevention in the elderly. The current study aimed to develop recurrent neural network models for extracting balance variables from a single inertial measurement unit (IMU) placed on the sacrum during walking. Thirteen healthy young and thirteen healthy older adults wore the IMU during walking and the ground truth of the inclination angles (IA) of the center of pressure to the center of mass vector and their rates of changes (RCIA) were measured simultaneously. The IA, RCIA, and IMU data were used to train four models (uni-LSTM, bi-LSTM, uni-GRU, and bi-GRU), with 10% of the data reserved to evaluate the model errors in terms of the root-mean-squared errors (RMSEs) and percentage relative RMSEs (rRMSEs). Independent <i>t</i>-tests were used for between-group comparisons. The sensitivity, specificity, and Pearson’s r for the effect sizes between the model-predicted data and experimental ground truth were also obtained. The bi-GRU with the weighted MSE model was found to have the highest prediction accuracy, computational efficiency, and the best ability in identifying statistical between-group differences when compared with the ground truth, which would be the best choice for the prolonged real-life monitoring of gait balance for fall risk management in the elderly. |
first_indexed | 2024-03-09T16:28:42Z |
format | Article |
id | doaj.art-b059b6f04b594ac4a219b5efde7aa385 |
institution | Directory Open Access Journal |
issn | 1424-8220 |
language | English |
last_indexed | 2024-03-09T16:28:42Z |
publishDate | 2023-11-01 |
publisher | MDPI AG |
record_format | Article |
series | Sensors |
spelling | doaj.art-b059b6f04b594ac4a219b5efde7aa3852023-11-24T15:05:06ZengMDPI AGSensors1424-82202023-11-012322904010.3390/s23229040Recurrent Neural Network Methods for Extracting Dynamic Balance Variables during Gait from a Single Inertial Measurement UnitCheng-Hao Yu0Chih-Ching Yeh1Yi-Fu Lu2Yi-Ling Lu3Ting-Ming Wang4Frank Yeong-Sung Lin5Tung-Wu Lu6Department of Biomedical Engineering, National Taiwan University, Taipei 10617, TaiwanDepartment of Biomedical Engineering, National Taiwan University, Taipei 10617, TaiwanDepartment of Information Management, National Taiwan University, Taipei 10617, TaiwanDepartment of Biomedical Engineering, National Taiwan University, Taipei 10617, TaiwanDepartment of Orthopaedic Surgery, School of Medicine, National Taiwan University, Taipei 10051, TaiwanDepartment of Information Management, National Taiwan University, Taipei 10617, TaiwanDepartment of Biomedical Engineering, National Taiwan University, Taipei 10617, TaiwanMonitoring dynamic balance during gait is critical for fall prevention in the elderly. The current study aimed to develop recurrent neural network models for extracting balance variables from a single inertial measurement unit (IMU) placed on the sacrum during walking. Thirteen healthy young and thirteen healthy older adults wore the IMU during walking and the ground truth of the inclination angles (IA) of the center of pressure to the center of mass vector and their rates of changes (RCIA) were measured simultaneously. The IA, RCIA, and IMU data were used to train four models (uni-LSTM, bi-LSTM, uni-GRU, and bi-GRU), with 10% of the data reserved to evaluate the model errors in terms of the root-mean-squared errors (RMSEs) and percentage relative RMSEs (rRMSEs). Independent <i>t</i>-tests were used for between-group comparisons. The sensitivity, specificity, and Pearson’s r for the effect sizes between the model-predicted data and experimental ground truth were also obtained. The bi-GRU with the weighted MSE model was found to have the highest prediction accuracy, computational efficiency, and the best ability in identifying statistical between-group differences when compared with the ground truth, which would be the best choice for the prolonged real-life monitoring of gait balance for fall risk management in the elderly.https://www.mdpi.com/1424-8220/23/22/9040balance controlrecurrent neural networkinertial measurement unitgait |
spellingShingle | Cheng-Hao Yu Chih-Ching Yeh Yi-Fu Lu Yi-Ling Lu Ting-Ming Wang Frank Yeong-Sung Lin Tung-Wu Lu Recurrent Neural Network Methods for Extracting Dynamic Balance Variables during Gait from a Single Inertial Measurement Unit Sensors balance control recurrent neural network inertial measurement unit gait |
title | Recurrent Neural Network Methods for Extracting Dynamic Balance Variables during Gait from a Single Inertial Measurement Unit |
title_full | Recurrent Neural Network Methods for Extracting Dynamic Balance Variables during Gait from a Single Inertial Measurement Unit |
title_fullStr | Recurrent Neural Network Methods for Extracting Dynamic Balance Variables during Gait from a Single Inertial Measurement Unit |
title_full_unstemmed | Recurrent Neural Network Methods for Extracting Dynamic Balance Variables during Gait from a Single Inertial Measurement Unit |
title_short | Recurrent Neural Network Methods for Extracting Dynamic Balance Variables during Gait from a Single Inertial Measurement Unit |
title_sort | recurrent neural network methods for extracting dynamic balance variables during gait from a single inertial measurement unit |
topic | balance control recurrent neural network inertial measurement unit gait |
url | https://www.mdpi.com/1424-8220/23/22/9040 |
work_keys_str_mv | AT chenghaoyu recurrentneuralnetworkmethodsforextractingdynamicbalancevariablesduringgaitfromasingleinertialmeasurementunit AT chihchingyeh recurrentneuralnetworkmethodsforextractingdynamicbalancevariablesduringgaitfromasingleinertialmeasurementunit AT yifulu recurrentneuralnetworkmethodsforextractingdynamicbalancevariablesduringgaitfromasingleinertialmeasurementunit AT yilinglu recurrentneuralnetworkmethodsforextractingdynamicbalancevariablesduringgaitfromasingleinertialmeasurementunit AT tingmingwang recurrentneuralnetworkmethodsforextractingdynamicbalancevariablesduringgaitfromasingleinertialmeasurementunit AT frankyeongsunglin recurrentneuralnetworkmethodsforextractingdynamicbalancevariablesduringgaitfromasingleinertialmeasurementunit AT tungwulu recurrentneuralnetworkmethodsforextractingdynamicbalancevariablesduringgaitfromasingleinertialmeasurementunit |