Use this URL to cite or link to this record in EThOS: https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.741040
Title: Smart technology for telerehabilitation : a smart device inertial-sensing method for gait analysis
Author: Steins, Dax
ISNI:       0000 0004 7230 6447
Awarding Body: Oxford Brookes University
Current Institution: Oxford Brookes University
Date of Award: 2016
Availability of Full Text:
Access from EThOS:
Access from Institution:
Abstract:
The aim of this work was to develop and validate an iPod Touch (4th generation) as a potential ambulatory monitoring system for clinical and non-clinical gait analysis. This thesis comprises four interrelated studies, the first overviews the current available literature on wearable accelerometry-based technology (AT) able to assess mobility-related functional activities in subjects with neurological conditions in home and community settings. The second study focuses on the detection of time-accurate and robust gait features from a single inertial measurement unit (IMU) on the lower back, establishing a reference framework in the process. The third study presents a simple step length algorithm for straight-line walking and the fourth and final study addresses the accuracy of an iPod’s inertial-sensing capabilities, more specifically, the validity of an inertial-sensing method (integrated in an iPod) to obtain time-accurate vertical lower trunk displacement measures. The systematic review revealed that present research primarily focuses on the development of accurate methods able to identify and distinguish different functional activities. While these are important aims, much of the conducted work remains in laboratory environments, with relatively little research moving from the “bench to the bedside.” This review only identified a few studies that explored AT’s potential outside of laboratory settings, indicating that clinical and real-world research significantly lags behind its engineering counterpart. In addition, AT methods are largely based on machine-learning algorithms that rely on a feature selection process. However, extracted features depend on the signal output being measured, which is seldom described. It is, therefore, difficult to determine the accuracy of AT methods without characterizing gait signals first. Furthermore, much variability exists among approaches (including the numbers of body-fixed sensors and sensor locations) to obtain useful data to analyze human movement. From an end-user’s perspective, reducing the amount of sensors to one instrument that is attached to a single location on the body would greatly simplify the design and use of the system. With this in mind, the accuracy of formerly identified or gait events from a single IMU attached to the lower trunk was explored. The study’s analysis of the trunk’s vertical and anterior-posterior acceleration pattern (and of their integrands) demonstrates, that a combination of both signals may provide more nuanced information regarding a person’s gait cycle, ultimately permitting more clinically relevant gait features to be extracted. Going one step further, a modified step length algorithm based on a pendulum model of the swing leg was proposed. By incorporating the trunk’s anterior-posterior displacement, more accurate predictions of mean step length can be made in healthy subjects at self-selected walking speeds. Experimental results indicate that the proposed algorithm estimates step length with errors less than 3% (mean error of 0.80 ± 2.01cm). The performance of this algorithm, however, still needs to be verified for those suffering from gait disturbances. Having established a referential framework for the extraction of temporal gait parameters as well as an algorithm for step length estimations from one instrument attached to the lower trunk, the fourth and final study explored the inertial-sensing capabilities of an iPod Touch. With the help of Dr. Ian Sheret and Oxford Brookes’ spin-off company ‘Wildknowledge’, a smart application for the iPod Touch was developed. The study results demonstrate that the proposed inertial-sensing method can reliably derive lower trunk vertical displacement (intraclass correlations ranging from .80 to .96) with similar agreement measurement levels to those gathered by a conventional inertial sensor (small systematic error of 2.2mm and a typical error of 3mm). By incorporating the aforementioned methods, an iPod Touch can potentially serve as a novel ambulatory monitor system capable of assessing gait in clinical and non-clinical environments.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.741040  DOI: Not available
Share: