Nowadays Inertial Measurement Unit (IMU) based systems are an alternative to the passive marker-based MOtion CAPture (MOCAP) procedure for clinical movement analysis. Although IMU-based motion capture is minimally invasive, faster, cheaper, and more versatile than MOCAP, a straightforward and immediate visualization of the data recorded via customized IMUs remains challenging. The first phase of the thesis aims to develop an algorithm for rapid and precise analysis and visualisation of body kinematics, using OpenSim software driven by IMUs prototypes, developed by Tallinn University of Technology, called "TinyTag". This will help to bridge the gap of visualisation lack for experimental IMUs. The algorithm is validated by comparing kinematic estimates, of four simple exercises (squat, frontal leg swing, knee to chest and heel to hamstring) performed by two healthy young adults (Age: 24 (±0) years; Height: 176 (±4) cm; Weight: 72.5 (±7)Kg), obtained from OpenSense driven by TinyTag against the MOCAP system. An additional comparison is made with the results obtained by OpenSense (OpenSim toolkit) driven by commercial Xsens IMUs. The study demonstrated the reliability of the kinematic variables processed by OpenSense using TinyTag data by comparing them with those obtained from the MOCAP system with an RMSE below the eligibility threshold of 5° for the Squat and Knee to Chest exercises. The bias of the difference between the two techniques is less than 10° for all exercises taken into analysis. In addition, Spearman's and Intraclass correlation coefficients demonstrated a perfect monotone correlation between the obtained results. Future studies could enlarge the population acquired and focus on reducing even more the difference with the gold standard. Furthermore, in the second phase of this work, the application of the developed algorithm in a water environment, exploiting the water-resistant feature of TinyTag sensors, is discussed. It is well known that there are benefits of physical activity in an aquatic setting, including buoyancy, hydrostatic pressure, and temperature. The use of water is indeed applied in sports and clinical fields for recovery from fatigue and treatment of chronic conditions, but no quantitative data are present for the exercises performed underwater, therefore is not possible to know whether the subject is correctly performing the exercise. In the second phase of the study, the aim is to define the physiological range of motion (ROM) for a distinct young healthy adult population performing four simple exercises commonly used in aquatic physical therapy. Twenty-five healthy young adult participants (Age: 22.4 ±1.7 years; Height: 176 ±7.8 cm; Weight: 70.16 ±10.4 Kg) are acquired while performing four exercises in two environments, on land and underwater. The joint angles of the subjects are computed starting from the data of experimental IMUs, processed by OpenSense. The results demonstrated that the ROM underwater is larger than the one on land in every exercise, except for the Squat. The latter show, negligible differences as the participants, due to the non-adjustable height of the water, did not perform the required exercise to their highest potential. The obtained results could be used as a baseline for comparison with pathological subjects or to follow the progress in a rehabilitation process. Future studies could incorporate the possibility of recruiting more people and using a pool whose height is adjustable.
Oggi i sistemi basati su unità di misura inerziale (IMU) sono un'alternativa alla procedura di MOtion CAPture (MOCAP) basata su marcatori passivi per l'analisi clinica del movimento. Sebbene la cattura del movimento basata su IMU sia minimamente invasiva, più veloce, più economica e più versatile della MOCAP, la visualizzazione immediata e diretta dei dati registrati tramite IMU personalizzate rimane una sfida. La prima fase della tesi mira a sviluppare un algoritmo per l'analisi e la visualizzazione rapida e precisa della cinematica corporea, utilizzando il software OpenSim guidato da prototipi di IMU, sviluppati dalla Tallinn University of Technology, chiamati "TinyTag". Questo aiuterà a colmare la mancanza di visualizzazione per le IMU sperimentali. L'algoritmo è stato validato confrontando le stime cinematiche di quattro semplici esercizi (squat, frontal leg swing, knee to chest and heel to hamstring) eseguiti da due giovani adulti sani (età: 24 (±0) anni; altezza: 176 (±4) cm; peso: 72,5 (±7)Kg), ottenute da OpenSense guidato dai dati dei TinyTag, con il sistema MOCAP. Un ulteriore confronto viene fatto con i risultati ottenuti da OpenSense (OpenSim toolkit) pilotato da IMU commerciali Xsens. Lo studio ha dimostrato l'affidabilità delle variabili cinematiche elaborate da OpenSense utilizzando i dati TinyTag, confrontandole con quelle ottenute dal sistema MOCAP con un RMSE inferiore alla soglia di ammissibilità di 5° per gli esercizi Squat e Knee to Chest. Il bias della differenza tra le due tecniche è inferiore a 10° per tutti gli esercizi presi in analisi. Inoltre, i coefficienti di correlazione di Spearman e Intraclasse hanno dimostrato una perfetta correlazione monotona tra i risultati ottenuti. Studi futuri potrebbero ampliare la popolazione acquisita e concentrarsi sulla riduzione della differenza con il gold standard. Nella seconda fase, viene discussa l'applicazione dell'algoritmo sviluppato in un ambiente acquatico, sfruttando la caratteristica di resistenza all'acqua dei sensori TinyTag. È noto che l'attività fisica in un ambiente acquatico presenta dei vantaggi, tra cui: il galleggiamento, la pressione idrostatica e la temperatura. L'uso dell'acqua è infatti comune in ambito sportivo e clinico per il recupero dalla fatica e il trattamento di patologie croniche, ma non sono presenti dati quantitativi per gli esercizi eseguiti sott'acqua, quindi non è possibile sapere se il soggetto sta eseguendo correttamente l'esercizio. Nella seconda fase dello studio, l'obiettivo è quello di definire il range fisiologico di movimento (ROM) per una popolazione distinta di giovani adulti sani che eseguono quattro semplici esercizi comunemente utilizzati nella terapia fisica acquatica. Venticinque partecipanti giovani adulti sani (età: 22.4 ±1.7 anni; altezza: 176 ±7.8 cm; peso: 70.16 ±10.4 Kg) sono stati ripresi mentre eseguivano quattro esercizi in due ambienti, a terra e sott'acqua. Gli angoli articolari dei soggetti sono calcolati a partire dai dati delle IMU sperimentali ed elaborati da OpenSense. I risultati hanno dimostrato che il ROM sott'acqua è maggiore di quello a terra in ogni esercizio, tranne che per lo Squat. Quest'ultimo presenta differenze trascurabili in quanto i partecipanti, a causa dell'altezza non regolabile dell'acqua, non hanno eseguito l'esercizio richiesto al massimo delle loro potenzialità. I risultati ottenuti potrebbero essere utilizzati come baseline per il confronto con soggetti patologici o per seguire i progressi di un processo di riabilitazione. Studi futuri potrebbero prevedere la possibilità di reclutare un maggior numero di persone e di utilizzare una piscina ad altezza regolabile.
Validation of IMU-based simulation for water exercises kinematics
Chieffo, Cristina;Motta, Filippo
2021/2022
Abstract
Nowadays Inertial Measurement Unit (IMU) based systems are an alternative to the passive marker-based MOtion CAPture (MOCAP) procedure for clinical movement analysis. Although IMU-based motion capture is minimally invasive, faster, cheaper, and more versatile than MOCAP, a straightforward and immediate visualization of the data recorded via customized IMUs remains challenging. The first phase of the thesis aims to develop an algorithm for rapid and precise analysis and visualisation of body kinematics, using OpenSim software driven by IMUs prototypes, developed by Tallinn University of Technology, called "TinyTag". This will help to bridge the gap of visualisation lack for experimental IMUs. The algorithm is validated by comparing kinematic estimates, of four simple exercises (squat, frontal leg swing, knee to chest and heel to hamstring) performed by two healthy young adults (Age: 24 (±0) years; Height: 176 (±4) cm; Weight: 72.5 (±7)Kg), obtained from OpenSense driven by TinyTag against the MOCAP system. An additional comparison is made with the results obtained by OpenSense (OpenSim toolkit) driven by commercial Xsens IMUs. The study demonstrated the reliability of the kinematic variables processed by OpenSense using TinyTag data by comparing them with those obtained from the MOCAP system with an RMSE below the eligibility threshold of 5° for the Squat and Knee to Chest exercises. The bias of the difference between the two techniques is less than 10° for all exercises taken into analysis. In addition, Spearman's and Intraclass correlation coefficients demonstrated a perfect monotone correlation between the obtained results. Future studies could enlarge the population acquired and focus on reducing even more the difference with the gold standard. Furthermore, in the second phase of this work, the application of the developed algorithm in a water environment, exploiting the water-resistant feature of TinyTag sensors, is discussed. It is well known that there are benefits of physical activity in an aquatic setting, including buoyancy, hydrostatic pressure, and temperature. The use of water is indeed applied in sports and clinical fields for recovery from fatigue and treatment of chronic conditions, but no quantitative data are present for the exercises performed underwater, therefore is not possible to know whether the subject is correctly performing the exercise. In the second phase of the study, the aim is to define the physiological range of motion (ROM) for a distinct young healthy adult population performing four simple exercises commonly used in aquatic physical therapy. Twenty-five healthy young adult participants (Age: 22.4 ±1.7 years; Height: 176 ±7.8 cm; Weight: 70.16 ±10.4 Kg) are acquired while performing four exercises in two environments, on land and underwater. The joint angles of the subjects are computed starting from the data of experimental IMUs, processed by OpenSense. The results demonstrated that the ROM underwater is larger than the one on land in every exercise, except for the Squat. The latter show, negligible differences as the participants, due to the non-adjustable height of the water, did not perform the required exercise to their highest potential. The obtained results could be used as a baseline for comparison with pathological subjects or to follow the progress in a rehabilitation process. Future studies could incorporate the possibility of recruiting more people and using a pool whose height is adjustable.| File | Dimensione | Formato | |
|---|---|---|---|
|
MSc_Thesis_Chieffo_Motta.pdf
accessibile in internet per tutti
Descrizione: MSc_Thesis
Dimensione
28.06 MB
Formato
Adobe PDF
|
28.06 MB | Adobe PDF | Visualizza/Apri |
|
Executive_Summary___Chieffo_Motta.pdf
accessibile in internet per tutti
Descrizione: Executive_Summary
Dimensione
11.96 MB
Formato
Adobe PDF
|
11.96 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/10589/203955