A global process of population ageing emerges from the latest demographic studies. Among the consequences of this phenomenon there is the growth of health care services users, with effects on both the quality of the delivered services and the financial costs to be sustained. During last decade many studies aimed to improve the efficiency of services for fragile people (e.g. disabled or elderly). Those studies aimed at developing Smart Environments providing Ambient Assisted Living (AAL) functionalities. Such systems leverage unobtrusive technologies, enabling the monitoring of domestic daily activities of fragile people and support their needs, guaranteeing them an independent life and reducing the need of intervention. This topic, thanks to Assistive Technology Group (ATG), is addressed by Politecnico di Milano with the research project named BRIDGe (Behaviour dRift compensation for autonomous and InDependent livinG). Few projects in the literature suggest measures useful to evaluate the routine of a person trough devices adopted in AAL systems. This thesis aims at providing a solution, live processing of data incoming from sensors installed in an elderly person's house. Leveraging Esper and Event Processing Language (EPL) it was possible to develop quantitative indicators based on simulated as well as on real data streams. The extracted indices are five and they concern various information about routine of a person: from basic aspects like (a) daily hours of activity measured by a single sensor, and (b) daily consumption of domestic services (water, electricity, ...) to higher level aspects, e.g. (c) identifying the occurrences of a complex activity, (d) evaluating the activeness of a single sensor and (e) the lifestyle of the monitored person considering all the sensors. This enables the extraction of information about different aspects of domestic habits and routines of elderly people, and applied to test datasets these indicators show to be effecting in describing the ordinary life conditions while enhancing behavior variations.
Recenti analisi demografiche mostrano che la vita media della popolazione mondiale si stia allungando. Questo può comportare una progressiva estensione del bacino di utenza dei servizi socio sanitari, con ripercussioni su qualità delle prestazioni erogate e quantità di risorse economiche necessarie. Molti lavori di ricerca sono stati svolti al fine di alleggerire il servizio sanitario dalle crescenti richieste di intervento delle fasce più sensibili. Tali studi si sono indirizzati verso lo sviluppo di sistemi che, attraverso funzionalità tipiche degli Smart Environment, fossero in grado di operare come Ambient Assisted Living (AAL). Questi, impiegando tecnologie non invasive, consentono il monitoraggio delle attività quotidiane degli anziani nell’ambiente domestico e di supportare le loro necessità garantendo un prolungato stile di vita indipendente. Il Politecnico di Milano, tramite l’Assistive Technology Group (ATG), sta sviluppando BRIDGe, un sistema atto a questo compito. La letteratura suggerisce alcune misure per valutare la routine di una persona che risiede in un AAL. Questo lavoro tesi si prefigge di fornire una risposta al problema sfruttando l’elaborazione in tempo reale dei dati provenienti dai dispositivi (sensori) installati nell’abitazione della persona assistita. Impiegando il linguaggio EPL ed Esper sono sviluppati indicatori quantitativi che, da flussi di dati reali (ARAS) o simulati (SHARON), costruiscono cinque stime utili per valutare lo stato di salute e di benessere di un soggetto. Sono individuati aspetti immediati come (a) le ore giornaliere di attività rilevate da un singolo sensore, e (b) i consumi giornalieri di servizi domestici. Successivamente aspetti più astratti come (c) il riconoscimento di un’attività complessa, (d) la valutazione dello stile di vita valutando entropia oraria e giornaliera relativa ad un singolo sensore nonché (e) quella relativa all’intera casa. Così si estraggono tramite indicatori quantitativi informazioni riguardo diversi aspetti delle abitudini casalinghe di una persona che, applicate su dataset di test, si dimostrano adeguate per descrivere le condizioni di vita ordinarie ed evidenziare eventuali anomalie.
Ambient assisted living : indicatori quantitativi per monitorare la routine di una persona
MOTTA, TOMMASO
2014/2015
Abstract
A global process of population ageing emerges from the latest demographic studies. Among the consequences of this phenomenon there is the growth of health care services users, with effects on both the quality of the delivered services and the financial costs to be sustained. During last decade many studies aimed to improve the efficiency of services for fragile people (e.g. disabled or elderly). Those studies aimed at developing Smart Environments providing Ambient Assisted Living (AAL) functionalities. Such systems leverage unobtrusive technologies, enabling the monitoring of domestic daily activities of fragile people and support their needs, guaranteeing them an independent life and reducing the need of intervention. This topic, thanks to Assistive Technology Group (ATG), is addressed by Politecnico di Milano with the research project named BRIDGe (Behaviour dRift compensation for autonomous and InDependent livinG). Few projects in the literature suggest measures useful to evaluate the routine of a person trough devices adopted in AAL systems. This thesis aims at providing a solution, live processing of data incoming from sensors installed in an elderly person's house. Leveraging Esper and Event Processing Language (EPL) it was possible to develop quantitative indicators based on simulated as well as on real data streams. The extracted indices are five and they concern various information about routine of a person: from basic aspects like (a) daily hours of activity measured by a single sensor, and (b) daily consumption of domestic services (water, electricity, ...) to higher level aspects, e.g. (c) identifying the occurrences of a complex activity, (d) evaluating the activeness of a single sensor and (e) the lifestyle of the monitored person considering all the sensors. This enables the extraction of information about different aspects of domestic habits and routines of elderly people, and applied to test datasets these indicators show to be effecting in describing the ordinary life conditions while enhancing behavior variations.| File | Dimensione | Formato | |
|---|---|---|---|
|
2016_04_Motta.pdf
accessibile in internet per tutti
Descrizione: Testo della tesi
Dimensione
4.02 MB
Formato
Adobe PDF
|
4.02 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/119301