Identification of gait disparity using gait analysis and cyclograms: application for prosthetic knee control / Nor Elleeiana Mohd Syah

Nor Elleeiana, Mohd Syah (2013) Identification of gait disparity using gait analysis and cyclograms: application for prosthetic knee control / Nor Elleeiana Mohd Syah. Masters thesis, University of Malaya.

Preview

PDF (Thesis M.A) Download (3252Kb) | Preview

Abstract

Choosing cyclograms over a function of time plots is very useful as a reference for designing a microprocessor-controlled prosthesis. This is because, the fact that locomotion is a tightly coordinated movement of several limb segments can be more naturally grasped as the coupled variables of two or more joints rather than from the study of individual joint kinematics or kinetics. Therefore, a new method developed by using the cyclic representation of locomotion and two-variable interaction, i.e. cyclograms, to determine the functional control parameter in terms of each gait sub-phases for prosthetic knee control design, is presented in this study. This study was divided into 3 phases. Phase 1 involved the gait data collection of normal subjects, transtibial amputees, transfemoral amputees and subjects wearing orthoses by using instrumented motion analysis system to characterize walking gait cyclograms. Phase 2 is where the cyclograms for the transtibial and transfemoral amputees were analyzed by using the neural network (NN) individually. The feedforward-backpropagation NN was used to predict each amputee’s gait characteristics from 10 trials and the deviations were determined at each gait sub-phase for each of the amputees. The results obtained from Phase 2 were then transferred to Phase 3, where the values at the sub-phase of the cyclograms that has greatest deviation percentage were extracted and replaced with the normal values and analyzed it again using NN. This study concluded that, in order to define the control design for the prosthetic knee, one should control at each gait sub-phase: Loading response – knee power, Mid-stance – Knee power, Terminal-stance – knee moment and knee angle, Pre-swing – Knee angle and knee power, Initial-swing – knee angle, Mid-swing – independent (no control), and terminal-swing – knee power. By controlling the knee parameter with respect to its affected joint kinetics and kinematics using artificial intelligence such as neural network, more natural gait can be achieved towards application in the prosthetic leg control design.

Actions (For repository staff only : Login required)