Queen's University - Utility Bar

QSpace at Queen's University >
Theses, Dissertations & Graduate Projects >
Queen's Theses & Dissertations >

Please use this identifier to cite or link to this item: http://hdl.handle.net/1974/6664

Title: Inertial Sensors in Estimating Spatio-Temporal Parameters of Walking: Performance Evaluation and Error Analysis
Authors: YANG, SHUOZHI

Files in This Item:

File Description SizeFormat
Yang_Shuozhi_201108_Master.pdf1.43 MBAdobe PDFView/Open
Keywords: stroke
gait analysis
biomechanics
kinematics
inertial sensor
gait asymmetry
walking speed
Issue Date: 23-Aug-2011
Series/Report no.: Canadian theses
Abstract: The portability, ease of use and improved accuracy of miniature inertial sensors brought by current microelectromechanical system (MEMS) technology has inspired researchers to develop human movement monitoring system with body-fixed sensors. Although a large number of studies have attempted to explore the use of miniature inertial sensors in estimating walking speed for the past two decades, there still remain some questions regarding applying inertial sensors in estimating walking speed under different walking conditions and for different subject populations. In this thesis, I focus on evaluating and improving the performance of a shank-mounted mounted inertial measurement unit (IMU) based walking speed estimation method. My research can be divided into four parts. The first part was a systematic review regarding the state of the art of current development of the inertial sensor based walking speed estimation method. A total of 16 articles were fully reviewed in terms of sensor specification, sensor attachment location, experimental design and spatial parameter estimation algorithm. In the second part, a comprehensive performance evaluation was conducted, which included the treadmill and overground walking experiments with constraint on the walking speed, stride length and stride frequency. A systematic error was observed in the error analysis of this study, which was adjusted by subtracting the bias by linear regression. In the third part, a post-stroke subject overground walking experiment was carried out with an improved walking speed estimation method that reduced the systematic error caused by previous false initial speed assumption. In addition to walking speed estimation, the gait asymmetry for post-stroke hemiparetic gait was also evaluated with the proposed method. The last part was the sensor error model analysis. We elaborately analyzed and discussed the estimation errors involved in this method in order to completely understand the sensor error compensation in walking speed estimation algorithm design. Two existing sensor error models and one newly developed sensor error model were compared with the treadmill walking experiment, which demonstrated the effect of each sensor error component on the estimation result and the importance of the sensor error model selection.
Description: Thesis (Master, Mechanical and Materials Engineering) -- Queen's University, 2011-08-23 19:38:16.965
URI: http://hdl.handle.net/1974/6664
Appears in Collections:Queen's Theses & Dissertations
Mechanical and Materials Engineering Graduate Theses

Items in QSpace are protected by copyright, with all rights reserved, unless otherwise indicated.

 

  DSpace Software Copyright © 2002-2008  The DSpace Foundation - TOP