Hybrid Spectral Estimation And Supervised Classification Techniques For Real Time Surface Electromyography (SEMG)

Kasno, Mohammad Afif (2018) Hybrid Spectral Estimation And Supervised Classification Techniques For Real Time Surface Electromyography (SEMG). Project Report. Universiti Teknikal Malaysia Melaka, Melaka, Malaysia. (Submitted)

Text (24 Pages) Hybrid Spectral Estimation And Supervised Classification Techniques For Real Time Surface Electromyography (SEMG).pdf - Submitted Version Restricted to Registered users only Download (658kB)

Abstract

Research on sEMG pattern recognition has attracted more and more attention for its application in rehabilitation and human-computer interfaces. However there are some recognized noises which will affect sEMG signals like inherent noise in electronics equipment,ambient noise, motion artifact and inherent instability of signal and some classified factors like causative,intermediate and deterministic factors which will affect sEMG signals that affect the accuracy of the pattern.The objective of this research is to investigate the characteristics of sEMG features using spectral estimation technique.Then,this study aims to assess the recognition rate of the sEMG pattern based on supervised classification algorithm. Finally this study intends to formulate a well-defined relationship between the spectral estimation and supervised classification technique to achieve the best trade-off between the system’s performance,computation time and hardware limitation in real time application.The study start by investigating the characteristics of sEMG features using spectral estimation technique. Subsequently,the recognition rate of the sEMG pattern based on supervised classification algorithm is assessed. Then,the relationship between the spectral estimation and supervised classification technique is formulated.Finally the performance of the pattern recognition technique is verified in real time application.It is expected that high optimized pattern recognition could be achieved considering the best trade-off between the system’s performance,computation time and hardware limitation in real time It could be extended to solve issues in other applications such as neuromuscular,kinesiology and prosthetic device in medical engineering field. Furthermore,it also could be applied for posture analysis and assistant device on high accuracy system in manufacturing field as well as human-computer interface and control device in electrical and electronic engineering.

Actions (login required)

View Item