Inertial measurement units (IMUs) have been developed tremendously and implemented to motion analysis widely nowadays, however the accuracy of wearable system is still the main issue due to integrated noise signal and inherent bias, even though sensor fusion is introduced to prevent the drift. This thesis purpose is to evaluate innovative full body wearable garment system based on IMUs against the reference standard system, optical motion capture system, in laboratory, through eight different sport exercises (lunge, squat, crunch, bicycle crunch, Russian twist, mountain climber, push-up and jumping jack), during which related angular kinematics like elbow flexion, knee flexion, arm abduction/adduction, trunk rotation, hip flexion/extension, hip abduction/adduction were measured accordingly. The IMU was embodied in the upper and lower body garments, while optical retroreflective markers were attached manually on the surface of the garments before subjects’ performance. The evaluation was accomplished by 12 healthy subjects. The joint kinematic data was recorded by both the optoelectronic system and IMUs system simultaneously. We compared both IMUs raw data and IMUS calibrated data to landmark system with statistics analysis like Bland-Altman, correlation, and found for most joint angles movement relative difference between IMU system and optical system was within 10° after calibration and high correlation coefficient(r>0.9) except elbow and shoulder (error introduce by incorrect post of arm during calibration). Main reason is location of the sensor for the thigh segment is placed on anterior part where rectus femoris could introduce error, as well as it could be impacted by skin artifact during moving. Nevertheless, this wearable system could provide useful reference motion measurement data, furthermore, combining 5G which could provide real-time coach, moreover, with ECG signal from the garment offering promising application in sport and clinical area.
Le unità di misura inerziali (IMU) si sono sviluppate moltissimo e vengono ampiamente utilizzate oggigiorno per l'analisi motoria. Tuttavia l'accuratezza dei sistemi portabili è ancora il problema principale a causa del rumore intrinseco, nonostante l'introduzione di tecniche quali la sensor fusion per evitare il drift. Lo scopo di questa tesi è valutare un sistema innovativo di indumenti dotato di IMU rispetto alla tecnica di riferimento tradizionale, basata su apparecchiatura ottica in laboratorio. Sono stati considerati otto diversi esercizi sportivi (balzi, squat, crunch, crunch bicicletta, twist russo, arrampicatore, push-ups e jumping jacks) durante i quali sono state misurate grandezze cinematiche angolari come la flessione dei gomiti, adduzione abduzione delle braccia, rotazione del busto, flessione estensione delle anche, adduzione abduzione delle anche e flessione delle ginocchia. Le IMU erano incorporate nelle parti inferiori e superiori dell’indumento , mentre dei markers ottici retroriflettenti sono stati applicati manualmente sulla superficie dell’ indumento prima dello svolgimento. Sono stati valutati dodici soggetti sani. I dati cinematici sopra menzionati sono stati registrati simultaneamente dall'apparecchiatura optoelettronica e dalle IMU. Abbiamo comparato sia i dati delle IMU sia la loro versione calibrata con quelli della tecnica tradizionale, eseguendo analisi statistiche quali grafici di Bland-Altman, correlazioni e trovato per la maggior parte delle misure angolari una differenza entro i 10° tra i dati IMU dopo la calibrazione e quelli ottici, e un alto coefficiente di correlazione (r<0.9) a parte per il gomito e la spalla (errore introdotto da una incorretta posa del braccio durante la calibrazione). La ragione principale di questa differenza è che la posizione del sensore per il segmento della coscia è situata nella parte anteriore dove il retto femorale potrebbe introdurre un errore, o anche degli artefatti introdotti dalla pelle durante il movimento. Nonostante ciò, questo sistema indossabile può fornire come riferimento utili misure del movimento. Inoltre, in combinazione con il 5G può fornire indicazioni per lo sport in real-time; infine, con il segnale ECG misurato con l’indumento, potrebbe costituire una promettente applicazione nello sport e in ambito medico.
Evaluation of a wearable system for motion analysis during different exercises
ZHANG, QIAN
2018/2019
Abstract
Inertial measurement units (IMUs) have been developed tremendously and implemented to motion analysis widely nowadays, however the accuracy of wearable system is still the main issue due to integrated noise signal and inherent bias, even though sensor fusion is introduced to prevent the drift. This thesis purpose is to evaluate innovative full body wearable garment system based on IMUs against the reference standard system, optical motion capture system, in laboratory, through eight different sport exercises (lunge, squat, crunch, bicycle crunch, Russian twist, mountain climber, push-up and jumping jack), during which related angular kinematics like elbow flexion, knee flexion, arm abduction/adduction, trunk rotation, hip flexion/extension, hip abduction/adduction were measured accordingly. The IMU was embodied in the upper and lower body garments, while optical retroreflective markers were attached manually on the surface of the garments before subjects’ performance. The evaluation was accomplished by 12 healthy subjects. The joint kinematic data was recorded by both the optoelectronic system and IMUs system simultaneously. We compared both IMUs raw data and IMUS calibrated data to landmark system with statistics analysis like Bland-Altman, correlation, and found for most joint angles movement relative difference between IMU system and optical system was within 10° after calibration and high correlation coefficient(r>0.9) except elbow and shoulder (error introduce by incorrect post of arm during calibration). Main reason is location of the sensor for the thigh segment is placed on anterior part where rectus femoris could introduce error, as well as it could be impacted by skin artifact during moving. Nevertheless, this wearable system could provide useful reference motion measurement data, furthermore, combining 5G which could provide real-time coach, moreover, with ECG signal from the garment offering promising application in sport and clinical area.| File | Dimensione | Formato | |
|---|---|---|---|
|
Evaluation of A Wearable System for Motion Analysis During Difference Exercises.pdf
solo utenti autorizzati dal 12/07/2020
Descrizione: Thesis text
Dimensione
5.13 MB
Formato
Adobe PDF
|
5.13 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/10589/149029