Automatically Segmenting Physical Performance Test Items for Older Adults Using a Doppler Radar: A Proof of Concept Study

Assessing the performance of physical activities through the modified physical performance test (mPPT) is a known approach for predicting frailty levels in older adults. This study proposes a system comprising a continuous-wave (CW) radar for data acquisition and deep neural network (DNN) models (co...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Yiyuan Zhang, Oluwatosin John Babarinde, Pengxuan Han, Xiangyu Wang, Peter Karsmakers, Dominique M. M.-P. Schreurs, Sabine Verschueren, Bart Vanrumste
Formato:	article
Lenguaje:	EN
Publicado:	IEEE 2021
Materias:	Activity performance monitoring deep learning frailty level older adults radar sensor Electrical engineering. Electronics. Nuclear engineering TK1-9971
Acceso en línea:	https://doaj.org/article/fcac2957838d4b72beaeb5aed3ff7578
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:fcac2957838d4b72beaeb5aed3ff7578
record_format	dspace
spelling	oai:doaj.org-article:fcac2957838d4b72beaeb5aed3ff75782021-11-20T00:01:53ZAutomatically Segmenting Physical Performance Test Items for Older Adults Using a Doppler Radar: A Proof of Concept Study2169-353610.1109/ACCESS.2021.3127327https://doaj.org/article/fcac2957838d4b72beaeb5aed3ff75782021-01-01T00:00:00Zhttps://ieeexplore.ieee.org/document/9611241/https://doaj.org/toc/2169-3536Assessing the performance of physical activities through the modified physical performance test (mPPT) is a known approach for predicting frailty levels in older adults. This study proposes a system comprising a continuous-wave (CW) radar for data acquisition and deep neural network (DNN) models (convolutional neural network (CNN) and convolutional recurrent neural network (CRNN)) as classifiers to automatically segment the mPPT items. These two DNN models were trained and evaluated in a leave-one-participant-out (LOPO) cross-validation procedure with a transfer learning method. To segment the mPPT items during recording by the radar, an additional flag activity was employed, which involves having the participants wave their hands at the start of each activity. Compared to the CNN, the CRNN achieved better classification performance, with the f1-score ranging from 0.3445 (<italic>lifting a book</italic>) to 0.9509 (<italic>standing balance</italic>). The recognition result was then used to segment the time-series data and predict each item’s duration. The average absolute duration prediction error ranged from 0.78 s (<italic>standing balance</italic>) to 2.78 s (<italic>climbing stairs</italic>). This result implies that the system has the potential to automatically segment mPPT items. Future works will be focused on accomplishing all the evaluation criteria automatically, for example, the steadiness and continuity of steps while turning 360°, and improving the low classification result of some mPPT items, such as <italic>lifting a book</italic>.Yiyuan ZhangOluwatosin John BabarindePengxuan HanXiangyu WangPeter KarsmakersDominique M. M.-P. SchreursSabine VerschuerenBart VanrumsteIEEEarticleActivity performance monitoringdeep learningfrailty levelolder adultsradar sensorElectrical engineering. Electronics. Nuclear engineeringTK1-9971ENIEEE Access, Vol 9, Pp 152765-152779 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	Activity performance monitoring deep learning frailty level older adults radar sensor Electrical engineering. Electronics. Nuclear engineering TK1-9971
spellingShingle	Activity performance monitoring deep learning frailty level older adults radar sensor Electrical engineering. Electronics. Nuclear engineering TK1-9971 Yiyuan Zhang Oluwatosin John Babarinde Pengxuan Han Xiangyu Wang Peter Karsmakers Dominique M. M.-P. Schreurs Sabine Verschueren Bart Vanrumste Automatically Segmenting Physical Performance Test Items for Older Adults Using a Doppler Radar: A Proof of Concept Study
description	Assessing the performance of physical activities through the modified physical performance test (mPPT) is a known approach for predicting frailty levels in older adults. This study proposes a system comprising a continuous-wave (CW) radar for data acquisition and deep neural network (DNN) models (convolutional neural network (CNN) and convolutional recurrent neural network (CRNN)) as classifiers to automatically segment the mPPT items. These two DNN models were trained and evaluated in a leave-one-participant-out (LOPO) cross-validation procedure with a transfer learning method. To segment the mPPT items during recording by the radar, an additional flag activity was employed, which involves having the participants wave their hands at the start of each activity. Compared to the CNN, the CRNN achieved better classification performance, with the f1-score ranging from 0.3445 (<italic>lifting a book</italic>) to 0.9509 (<italic>standing balance</italic>). The recognition result was then used to segment the time-series data and predict each item’s duration. The average absolute duration prediction error ranged from 0.78 s (<italic>standing balance</italic>) to 2.78 s (<italic>climbing stairs</italic>). This result implies that the system has the potential to automatically segment mPPT items. Future works will be focused on accomplishing all the evaluation criteria automatically, for example, the steadiness and continuity of steps while turning 360°, and improving the low classification result of some mPPT items, such as <italic>lifting a book</italic>.
format	article
author	Yiyuan Zhang Oluwatosin John Babarinde Pengxuan Han Xiangyu Wang Peter Karsmakers Dominique M. M.-P. Schreurs Sabine Verschueren Bart Vanrumste
author_facet	Yiyuan Zhang Oluwatosin John Babarinde Pengxuan Han Xiangyu Wang Peter Karsmakers Dominique M. M.-P. Schreurs Sabine Verschueren Bart Vanrumste
author_sort	Yiyuan Zhang
title	Automatically Segmenting Physical Performance Test Items for Older Adults Using a Doppler Radar: A Proof of Concept Study
title_short	Automatically Segmenting Physical Performance Test Items for Older Adults Using a Doppler Radar: A Proof of Concept Study
title_full	Automatically Segmenting Physical Performance Test Items for Older Adults Using a Doppler Radar: A Proof of Concept Study
title_fullStr	Automatically Segmenting Physical Performance Test Items for Older Adults Using a Doppler Radar: A Proof of Concept Study
title_full_unstemmed	Automatically Segmenting Physical Performance Test Items for Older Adults Using a Doppler Radar: A Proof of Concept Study
title_sort	automatically segmenting physical performance test items for older adults using a doppler radar: a proof of concept study
publisher	IEEE
publishDate	2021
url	https://doaj.org/article/fcac2957838d4b72beaeb5aed3ff7578
work_keys_str_mv	AT yiyuanzhang automaticallysegmentingphysicalperformancetestitemsforolderadultsusingadopplerradaraproofofconceptstudy AT oluwatosinjohnbabarinde automaticallysegmentingphysicalperformancetestitemsforolderadultsusingadopplerradaraproofofconceptstudy AT pengxuanhan automaticallysegmentingphysicalperformancetestitemsforolderadultsusingadopplerradaraproofofconceptstudy AT xiangyuwang automaticallysegmentingphysicalperformancetestitemsforolderadultsusingadopplerradaraproofofconceptstudy AT peterkarsmakers automaticallysegmentingphysicalperformancetestitemsforolderadultsusingadopplerradaraproofofconceptstudy AT dominiquemmpschreurs automaticallysegmentingphysicalperformancetestitemsforolderadultsusingadopplerradaraproofofconceptstudy AT sabineverschueren automaticallysegmentingphysicalperformancetestitemsforolderadultsusingadopplerradaraproofofconceptstudy AT bartvanrumste automaticallysegmentingphysicalperformancetestitemsforolderadultsusingadopplerradaraproofofconceptstudy
_version_	1718419874652356608

Automatically Segmenting Physical Performance Test Items for Older Adults Using a Doppler Radar: A Proof of Concept Study

Ejemplares similares