Surface electromyography (SEMG) normalization method based on pre fatigue maximal voluntary contraction
Surface electromyography (sEMG) pattern recognition task requires high accuracy classification. However, current technology suffers from two main problems. The first problem is inconsistent pattern due to fatigue while the second is robustness of sEMG features due to low signal to noise ratio, SNR....
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English English |
| Published: |
2017
|
| Subjects: | |
| Online Access: | http://eprints.utem.edu.my/id/eprint/20616/1/Surface%20Electromyography%20%28SEMG%29%20Normalization%20Method%20Based%20On%20Pre%20Fatigue%20Maximal%20Voluntary%20Contraction.pdf http://eprints.utem.edu.my/id/eprint/20616/2/Surface%20electromyography%20%28SEMG%29%20normalization%20method%20based%20on%20pre%20fatigue%20maximal%20voluntary%20contraction.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Surface electromyography (sEMG) pattern recognition task requires high accuracy classification. However, current technology suffers from two main problems. The first problem is inconsistent pattern due to fatigue while the second is robustness of sEMG features due to low signal to noise ratio, SNR. This research intends to address both sEMG problems mentioned by proposing a normalization method named as pre fatigue maximal voluntary contraction (PFMVC) and a feature known as Maximal Amplitude Spectrum (MaxPS). The both method used to carry the objectives, the first is to analyse a normalization method based on pre fatigue maximal voluntary contraction (PFMVC) and the second objective is to design and verify a new features known as Maximal Amplitude Spectrum (MaxPS). It is found that the proposed method improves sEMG pattern recognition accuracy by 98.48% when compare to 97%. The performance of PFMVC normalization method is measured by mean variance of boxplot across several subject which is reduce inconsistency from 3.41x10-3 to 1.73x10-3 , p-value of one way analysis of variance (One- Way ANOVA) is reduce from p=0.25 to p=0.035 and variance of mean intra class correlation co-efficient, (ICC) is reduce from 26x10-4 to 7.089x10-4. The precision and robust of MaxPS features is determine by lowest Error to mean percentage (%ETM) which is 0.213 , lowest in Euclidean distance,(Ed) which is 0.0034 and lowest hoteling t2 which is 0.27. From the results, it shows that the MaxPS is a robust and precise feature for force and fatigue indicator. This will give the benefit for force and fatigue mapping application. |
|---|