Development of paraplegic quadriceps muscle models using artificial intelligent techniques / Syafinas Mohd Salleh

Mohd Salleh, Syafinas (2015) Development of paraplegic quadriceps muscle models using artificial intelligent techniques / Syafinas Mohd Salleh. Masters thesis, Universiti Teknologi MARA.

[img] Text
TM_SYAFINAS MOHD SALLEH EE 15_5.pdf

Download (1MB)

Abstract

This research presents the development of quadriceps muscle models using several Artificial Intelligent techniques. The data of paraplegic muscle behavior consists of frequency, sampling time, pulse width and muscle torque have been obtained from Hospital Sungai Buloh. The data was collected from 20 paraplegic patient’s age between 40-60 years olds throughout the study. 722 data obtained from the data collection to be used for muscle model development. In this study Artificial Neural Network (ANN) will be used to develop paraplegic muscle models. There are Levenberg Marquardt (LM),Resilient Backpropagation (RP) and Scale Conjugate Gradient (SCG). Other AI techniques involve in this study are Fuzzy Logic and Adaptive Neural Fuzzy Interference System (ANFIS). The last technique that has been selected for the muscle model is Extended Kalman Filter (EKF) techniques. From the result, LM gives the best performance compare to others model developed. From analysis for all model develops, LM gives the lowest Mean Squared Error (MSE) 0.2007 followed by RP with MSE 0.2644, SCG 0.3758, Fuzzy Logic 0.7557, ANFIS 1.36 and EKF 29.68. In this analysis,the proposed EKF shows unreliable results where it produces the highest error compared to the other techniques. It shows that EKF is not suitable to be used for develop quadriceps muscle model with unsynchronized type of data behavior. The result obtained in this study can be used in the future for design and evaluation of various control strategies that could help avoid injuries during the experiment on real paraplegic patients

Item Type: Thesis (Masters)
Creators:
CreatorsEmail
Mohd Salleh, SyafinasUNSPECIFIED
Subjects: Q Science > QA Mathematics > Instruments and machines > Electronic computers. Computer science > Artificial immune systems. Immunocomputers
Q Science > QA Mathematics > Instruments and machines > Electronic computers. Computer science > Artificial immune systems. Immunocomputers
Divisions: Faculty of Electrical Engineering
Item ID: 15276
Uncontrolled Keywords: Paraplegic Quadriceps Muscle Models; A,rtificial Intelligent Techniques
Last Modified: 22 Nov 2016 02:59
Depositing User: Staf Pendigitalan 2
URI: http://ir.uitm.edu.my/id/eprint/15276

Actions (login required)

View Item View Item

Downloads

Downloads per month over past year