Real-time motion onset recognition for robot-assisted gait rehabilitation.
Body weight support
EtherCAT
Inertial measurement unit
Real-time activity recognition
Sliding window
Wireless EtherCAT interface
Journal
Journal of neuroengineering and rehabilitation
ISSN: 1743-0003
Titre abrégé: J Neuroeng Rehabil
Pays: England
ID NLM: 101232233
Informations de publication
Date de publication:
28 01 2022
28 01 2022
Historique:
received:
27
05
2021
accepted:
29
12
2021
entrez:
29
1
2022
pubmed:
30
1
2022
medline:
22
3
2022
Statut:
epublish
Résumé
Many patients with neurological movement disorders fear to fall while performing postural transitions without assistance, which prevents them from participating in daily life. To overcome this limitation, multi-directional Body Weight Support (BWS) systems have been developed allowing them to perform training in a safe environment. In addition to overground walking, these innovative/novel systems can assist patients to train many more gait-related tasks needed for daily life under very realistic conditions. The necessary assistance during the users' movements can be provided via task-dependent support designs. One remaining challenge is the manual switching between task-dependent supports. It is error-prone, cumbersome, distracts therapists and patients, and interrupts the training workflow. Hence, we propose a real-time motion onset recognition model that performs automatic support switching between standing-up and sitting-down transitions and other gait-related tasks (8 classes in total). To predict the onsets of the gait-related tasks, three Inertial Measurement Units (IMUs) were attached to the sternum and middle of outer thighs of 19 controls without neurological movement disorders and two individuals with incomplete Spinal Cord Injury (iSCI). The data of IMUs obtained from different gait tasks was sent synchronously to a real-time data acquisition system through a custom-made Bluetooth-EtherCAT gateway. In the first step, data was applied offline for training five different classifiers. The best classifier was chosen based on F1-score results of a Leave-One-Participant-Out Cross-Validation (LOPOCV), which is an unbiased way of testing. In a final step, the chosen classifier was tested in real time with an additional control participant to demonstrate feasibility for real-time classification. Testing five different classifiers, the best performance was obtained in a single-layer neural network with 25 neurons. The F1-score of [Formula: see text] and [Formula: see text] are achieved on testing using LOPOCV and test data ([Formula: see text], participants = 20), respectively. Furthermore, the results from the implemented real-time classifier were compared with the offline classifier and revealed nearly identical performance (difference = [Formula: see text]). A neural network classifier was trained for identifying the onset of gait-related tasks in real time. Test data showed convincing performance for offline and real-time classification. This demonstrates the feasibility and potential for implementing real-time onset recognition in rehabilitation devices in future.
Sections du résumé
BACKGROUND
Many patients with neurological movement disorders fear to fall while performing postural transitions without assistance, which prevents them from participating in daily life. To overcome this limitation, multi-directional Body Weight Support (BWS) systems have been developed allowing them to perform training in a safe environment. In addition to overground walking, these innovative/novel systems can assist patients to train many more gait-related tasks needed for daily life under very realistic conditions. The necessary assistance during the users' movements can be provided via task-dependent support designs. One remaining challenge is the manual switching between task-dependent supports. It is error-prone, cumbersome, distracts therapists and patients, and interrupts the training workflow. Hence, we propose a real-time motion onset recognition model that performs automatic support switching between standing-up and sitting-down transitions and other gait-related tasks (8 classes in total).
METHODS
To predict the onsets of the gait-related tasks, three Inertial Measurement Units (IMUs) were attached to the sternum and middle of outer thighs of 19 controls without neurological movement disorders and two individuals with incomplete Spinal Cord Injury (iSCI). The data of IMUs obtained from different gait tasks was sent synchronously to a real-time data acquisition system through a custom-made Bluetooth-EtherCAT gateway. In the first step, data was applied offline for training five different classifiers. The best classifier was chosen based on F1-score results of a Leave-One-Participant-Out Cross-Validation (LOPOCV), which is an unbiased way of testing. In a final step, the chosen classifier was tested in real time with an additional control participant to demonstrate feasibility for real-time classification.
RESULTS
Testing five different classifiers, the best performance was obtained in a single-layer neural network with 25 neurons. The F1-score of [Formula: see text] and [Formula: see text] are achieved on testing using LOPOCV and test data ([Formula: see text], participants = 20), respectively. Furthermore, the results from the implemented real-time classifier were compared with the offline classifier and revealed nearly identical performance (difference = [Formula: see text]).
CONCLUSIONS
A neural network classifier was trained for identifying the onset of gait-related tasks in real time. Test data showed convincing performance for offline and real-time classification. This demonstrates the feasibility and potential for implementing real-time onset recognition in rehabilitation devices in future.
Identifiants
pubmed: 35090511
doi: 10.1186/s12984-022-00984-x
pii: 10.1186/s12984-022-00984-x
pmc: PMC8796576
doi:
Types de publication
Journal Article
Research Support, Non-U.S. Gov't
Langues
eng
Sous-ensembles de citation
IM
Pagination
11Informations de copyright
© 2022. The Author(s).
Références
IEEE Trans Inf Technol Biomed. 2006 Jan;10(1):156-67
pubmed: 16445260
Sensors (Basel). 2018 Aug 31;18(9):
pubmed: 30200377
J Biomech. 1991;24(1):77-85
pubmed: 2026635
IEEE Trans Biomed Eng. 2010 Mar;57(3):542-51
pubmed: 19846361
Pervasive Mob Comput. 2014 Feb 1;10(Pt B):138-154
pubmed: 24729780
Brain. 2004 Jun;127(Pt 6):1403-14
pubmed: 15069022
IEEE Int Conf Rehabil Robot. 2013 Jun;2013:6650512
pubmed: 24187327
Brain Res Bull. 2008 Jul 30;76(5):459-63
pubmed: 18534251
Spinal Cord. 2006 May;44(5):287-96
pubmed: 16186857
IEEE Trans Inf Technol Biomed. 2010 Sep;14(5):1166-72
pubmed: 20529753
Sensors (Basel). 2017 Nov 23;17(12):
pubmed: 29168769
IEEE Trans Neural Syst Rehabil Eng. 2006 Sep;14(3):311-21
pubmed: 17009491
Med Eng Phys. 2014 Jun;36(6):779-85
pubmed: 24636448
IEEE Trans Haptics. 2015 Oct-Dec;8(4):421-33
pubmed: 25974949
Am J Phys Med Rehabil. 2012 Nov;91(11 Suppl 3):S232-41
pubmed: 23080039
J Rehabil Res Dev. 2011;48(4):287-98
pubmed: 21674384
J Rehabil Med. 2010 Jun;42(6):513-9
pubmed: 20549154
Sensors (Basel). 2015 Dec 11;15(12):31314-38
pubmed: 26690450
IEEE Int Conf Rehabil Robot. 2017 Jul;2017:412-417
pubmed: 28813854
IEEE Trans Rehabil Eng. 2000 Dec;8(4):523-9
pubmed: 11204044
IEEE Trans Biomed Eng. 2009 Mar;56(3):871-9
pubmed: 19272902
Spinal Cord. 2006 Nov;44(11):668-75
pubmed: 16462822
IEEE Trans Neural Syst Rehabil Eng. 2008 Jun;16(3):286-97
pubmed: 18586608
J Spinal Cord Med. 2011 Nov;34(6):535-46
pubmed: 22330108
J Comp Neurol. 2004 Aug 16;476(2):130-45
pubmed: 15248194