Research Article
Elderly People Activity Recognition in Smart Grid
Monitoring Environment
Anusha Ganesan ,
1
Anand Paul ,
1
and HyunCheol Seo
2
1
School of Computer Science and Engineering Kyungpook National University, Daegu 41566, Republic of Korea
2
School of Architectural, Civil, Environmental and Energy Engineering, Kyungpook National University,
Daegu, Republic of Korea
CorrespondenceshouldbeaddressedtoHyunCheolSeo;charles@knu.ac.kr
Received 3 February 2022; Accepted 7 March 2022; Published 22 March 2022
AcademicEditor:RaviSamikannu
Copyright©2022AnushaGanesanetal.isisanopenaccessarticledistributedundertheCreativeCommonsAttribution
License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is
properlycited.
Elderlypeopleactivityrecognitionhasbecomeavitalnecessityinmanycountries,becausemostoftheelderlypeoplelivealone
andarevulnerable.us,moreresearchtoadvanceinthemonitoringsystemsusedtorecognizetheactivitiesofelderlypeopleis
required.Manyresearchershaveproposeddifferentmonitoringsystemsforactivityrecognitionusingwiredandwirelesswearable
sensingdevices.However,theactivityclassificationaccuracyachievedsofarshouldbeimprovedtomeetthechallengesofmore
preciseactivitymonitoring.OurstudyproposesasmartHumanActivityRecognitionsystemarchitectureutilizinganopensource
dataset generated by wireless, batteryless sensors used by 14 healthy aged persons and unsupervised and supervised machine
learning algorithms. In this paper, we also propose using a smart grid for checking regularly the wearable sensing device
operational status to address the well-known reliability challenges of these devices, such as wireless charging and data trust-
worthiness.Asthedatafromthesensingdeviceisverynoisy,weemploytheK-means++clusteringtoidentifyoutliersanduse
advancedensembleclassificationtechniques,suchasthestackingclassifierforwhichametamodelbuiltusingtherandomforest
algorithmgavebetterresultsthanallotherbasemodelsconsidered.Wealsoemployabaggingclassifier,whichisanensemble
meta-estimatorfittingthepredictionoutputsofthebaseclassifiersandaggregatingthemtoproducetheensembleoutput.ebest
classificationaccuracyof99.81wasachievedbythestackingclassifierintrainingand99.78%intesting,respectively.Comparisons
forfindingthebestmodelwereconductedusingtherecall,F1score,andprecisionvalues.
1. Introduction
Several countries in the world currently have a vast elderly
population. is situation entails additional challenges in
providing quality healthcare services and facilities for this
demographic. Elderly people require more physical and
psychological support and assistance. Most of the elderly
peoplecurrentlyliveontheirownastheirchildrenworkin
the different geographic locations. is leads to a lesser
likelihoodofchildrentakingcareoftheirparents.ismakes
theelderlyparentsveryvulnerabletoriskswithnoimmediate
assistanceavailable.However,thisproblemcanbemitigated
with the technological advancements in activity monitoring
systemsforelderlypeople.Datarequiredbythesesystemsare
acquired by wearable sensing devices and then analyzed to
understandandforecasttheindividual’shealthconditionand
requiredsupport.esemonitoringsystemsarereferredtoas
Human Activity Recognition (HAR) systems. HAR systems
play important roles in several domains such as healthcare,
security, and smart environment deployment [1]. e op-
erationofthesesystemsinvolvesfivemajorsteps,whichare
illustratedinFigure1forfourbasicactivitiessuchassleeping,
walking, standing, and sitting.
However, to successfully implement these steps, ap-
propriate devices and sensors are required to ensure the
efficiency of the entire system. Hence, the development of
such system relies on wireless networks, machine learning
(ML), data processing, and classification methods. A HAR
systemcandetectandmonitortheactivitiesaswellasthe
hazardsthatcanaffecttheelderlypeople.Sincethedevices
Hindawi
Mathematical Problems in Engineering
Volume 2022, Article ID 9540033, 12 pages
https://doi.org/10.1155/2022/9540033