Estimation of Lower Limb Kinematics during Squat Task in Different Loading Using sEMG Activity and Deep Recurrent Neural Networks
The aim of the present study was to predict the kinematics of the knee and the ankle joints during a squat training task of different intensities. Lower limb surface electromyographic (sEMG) signals and the 3-D kinematics of lower extremity joints were recorded from 19 body builders during squat tra...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2021-11-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1424-8220/21/23/7773 |
| _version_ | 1797507222652059648 |
|---|---|
| author | Alireza Rezaie Zangene Ali Abbasi Kianoush Nazarpour |
| author_facet | Alireza Rezaie Zangene Ali Abbasi Kianoush Nazarpour |
| author_sort | Alireza Rezaie Zangene |
| collection | DOAJ |
| description | The aim of the present study was to predict the kinematics of the knee and the ankle joints during a squat training task of different intensities. Lower limb surface electromyographic (sEMG) signals and the 3-D kinematics of lower extremity joints were recorded from 19 body builders during squat training at four loading conditions. A long-short term memory (LSTM) was used to estimate the kinematics of the knee and the ankle joints. The accuracy, in terms root-mean-square error (RMSE) metric, of the LSTM network for the knee and ankle joints were 6.774 ± 1.197 and 6.961 ± 1.200, respectively. The LSTM network with inputs processed by cross-correlation (CC) method showed 3.8% and 4.7% better performance in the knee and ankle joints, respectively, compared to when the CC method was not used. Our results showed that in the prediction, regardless of the intensity of movement and inter-subject variability, an off-the-shelf LSTM decoder outperforms conventional fully connected neural networks. |
| first_indexed | 2024-03-10T04:45:28Z |
| format | Article |
| id | doaj.art-739d78b8a24549158c24751a2e7e660a |
| institution | Directory Open Access Journal |
| issn | 1424-8220 |
| language | English |
| last_indexed | 2024-03-10T04:45:28Z |
| publishDate | 2021-11-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj.art-739d78b8a24549158c24751a2e7e660a2023-11-23T02:58:58ZengMDPI AGSensors1424-82202021-11-012123777310.3390/s21237773Estimation of Lower Limb Kinematics during Squat Task in Different Loading Using sEMG Activity and Deep Recurrent Neural NetworksAlireza Rezaie Zangene0Ali Abbasi1Kianoush Nazarpour2Department of Biomechanics and Sports Injuries, Faculty of Physical Education and Sports Sciences, Kharazmi University, Tehran 15719-14911, IranDepartment of Biomechanics and Sports Injuries, Faculty of Physical Education and Sports Sciences, Kharazmi University, Tehran 15719-14911, IranSchool of Informatics, The University of Edinburgh, Edinburgh EH8 9AB, UKThe aim of the present study was to predict the kinematics of the knee and the ankle joints during a squat training task of different intensities. Lower limb surface electromyographic (sEMG) signals and the 3-D kinematics of lower extremity joints were recorded from 19 body builders during squat training at four loading conditions. A long-short term memory (LSTM) was used to estimate the kinematics of the knee and the ankle joints. The accuracy, in terms root-mean-square error (RMSE) metric, of the LSTM network for the knee and ankle joints were 6.774 ± 1.197 and 6.961 ± 1.200, respectively. The LSTM network with inputs processed by cross-correlation (CC) method showed 3.8% and 4.7% better performance in the knee and ankle joints, respectively, compared to when the CC method was not used. Our results showed that in the prediction, regardless of the intensity of movement and inter-subject variability, an off-the-shelf LSTM decoder outperforms conventional fully connected neural networks.https://www.mdpi.com/1424-8220/21/23/7773surface electromyography (sEMG)continuous estimationdeep neural networks (DNNs)joint angle estimationsquat |
| spellingShingle | Alireza Rezaie Zangene Ali Abbasi Kianoush Nazarpour Estimation of Lower Limb Kinematics during Squat Task in Different Loading Using sEMG Activity and Deep Recurrent Neural Networks Sensors surface electromyography (sEMG) continuous estimation deep neural networks (DNNs) joint angle estimation squat |
| title | Estimation of Lower Limb Kinematics during Squat Task in Different Loading Using sEMG Activity and Deep Recurrent Neural Networks |
| title_full | Estimation of Lower Limb Kinematics during Squat Task in Different Loading Using sEMG Activity and Deep Recurrent Neural Networks |
| title_fullStr | Estimation of Lower Limb Kinematics during Squat Task in Different Loading Using sEMG Activity and Deep Recurrent Neural Networks |
| title_full_unstemmed | Estimation of Lower Limb Kinematics during Squat Task in Different Loading Using sEMG Activity and Deep Recurrent Neural Networks |
| title_short | Estimation of Lower Limb Kinematics during Squat Task in Different Loading Using sEMG Activity and Deep Recurrent Neural Networks |
| title_sort | estimation of lower limb kinematics during squat task in different loading using semg activity and deep recurrent neural networks |
| topic | surface electromyography (sEMG) continuous estimation deep neural networks (DNNs) joint angle estimation squat |
| url | https://www.mdpi.com/1424-8220/21/23/7773 |
| work_keys_str_mv | AT alirezarezaiezangene estimationoflowerlimbkinematicsduringsquattaskindifferentloadingusingsemgactivityanddeeprecurrentneuralnetworks AT aliabbasi estimationoflowerlimbkinematicsduringsquattaskindifferentloadingusingsemgactivityanddeeprecurrentneuralnetworks AT kianoushnazarpour estimationoflowerlimbkinematicsduringsquattaskindifferentloadingusingsemgactivityanddeeprecurrentneuralnetworks |