Manual wheelchair users often have an inactive lifestyle and hence it is easier for them to suffer from diabetes, obesity, cardiovascular problems and cancer. To prevent such inactive lifestyles, and thus indirectly help them to live a healthier life, the Norwegian University of Science and Technology (NTNU) has started working on the so-called ‘digital wheelchair project’. Its objective is developing and testing a wheelchair augmented with IT technologies that, through the help of algorithms that monitor health related parameters and opportune sensing hardware, may reliably estimate the daily energy expenditure of such wheelchair users (a feat that has always been a challenging one). This dissertation is performed in collaboration with NTNU - Norwegian University of Science and Technology - and aims at contributing to the modeling of the temporal dynamics of and correlations between heart rate and energy expenditures data. The work focuses thus on applying statistical learning and system identification concepts to reconstruct holistic (and individualized per-patient) individualized first order Ordinary Differential Equations from heart rate and energy expenditure field data. A first order model has proved to be too poor to describe such a complex system and the multitude of outliers requires a more robust approach than the simple Maximum Likelihood one, for instance one might consider to include priors. However, we have set the basis for further studies and research developments.
Le persone in carrozzina spesso hanno uno stile di vita poco attivo e quindi e più probabile che soffrano di diabete, obesità, problemi cardiovascolari e cancro. Per prevenire tale stile di vita poco attivo e pertanto aiutarli indirettamente a vivere una vita più sana, l’Università Norvegese della Scienza e Tecnologia (NTNU) ha iniziato a lavorare al cosiddetto ‘digital wheelchair project’. Tale progetto consiste nello sviluppo e prova di una sedia a rotelle potenziata da tecnologie IT che, attraverso l’aiuto di algoritmi che monitorano parametri collegati alla salute e di opportuni sistemi di rilevamento hardware, possano stimare in modo affidabile l’energia spesa proprio dalle persone in carrozzina (una sfida che è sempre stata impegnativa e che non ha ancora soluzione). Questa tesi e stata svolta in collaborazione con NTNU - Università Norvegese della Scienza e Tecnologia - con lo scopo di contribuire alla modellazione delle dinamiche temporali del battito cardiaco e del dispendio energetico e di analizzarne una possibile correlazione. Questo lavoro si focalizza sull’applicazione delle nozioni statistiche imparate e dei concetti di identificazione di sistemi per ricostruire Equazioni Differenziali Ordinarie olistiche (e personalizzate per paziente) a partire dai dati sperimentali di battito cardiaco e spesa energetica. Un modello del primo ordine si è rivelato troppo povero per descrivere un sistema così complesso e il vasto numero di dati da escludere richiede un approccio più robusto di uno semplice a massima verosimiglianza, per esempio si potrebbero includere delle conoscenze a priori. Ciononostante, abbiamo gettato le basi per futuri studi e sviluppi del progetto.
Modelling and identification of heart rate and energy expenditure dynamics for manual wheelchair users.
CAPPOZZO, VITTORIA
2021/2022
Abstract
Manual wheelchair users often have an inactive lifestyle and hence it is easier for them to suffer from diabetes, obesity, cardiovascular problems and cancer. To prevent such inactive lifestyles, and thus indirectly help them to live a healthier life, the Norwegian University of Science and Technology (NTNU) has started working on the so-called ‘digital wheelchair project’. Its objective is developing and testing a wheelchair augmented with IT technologies that, through the help of algorithms that monitor health related parameters and opportune sensing hardware, may reliably estimate the daily energy expenditure of such wheelchair users (a feat that has always been a challenging one). This dissertation is performed in collaboration with NTNU - Norwegian University of Science and Technology - and aims at contributing to the modeling of the temporal dynamics of and correlations between heart rate and energy expenditures data. The work focuses thus on applying statistical learning and system identification concepts to reconstruct holistic (and individualized per-patient) individualized first order Ordinary Differential Equations from heart rate and energy expenditure field data. A first order model has proved to be too poor to describe such a complex system and the multitude of outliers requires a more robust approach than the simple Maximum Likelihood one, for instance one might consider to include priors. However, we have set the basis for further studies and research developments.File | Dimensione | Formato | |
---|---|---|---|
Cappozzo_Vittoria.pdf
accesso aperto
Dimensione
5.23 MB
Formato
Adobe PDF
|
5.23 MB | Adobe PDF | Visualizza/Apri |
The text of this website © Università degli studi di Padova. Full Text are published under a non-exclusive license. Metadata are under a CC0 License
https://hdl.handle.net/20.500.12608/34626