ArticlePDF Available

Abstract

Electromyography (EMG) signal is non-stationary signal and highly complex time and frequency characteristics. Fast-Fourier transform common technique in signal processing involving EMG signal. However, this technique has a limitation to provide the time-frequency information for EMG signals. This paper presents the analysis of EMG signal of the variable lifting height and mass of load between the four subjects selected in manual lifting by using spectrogram. Spectrogram is one of the time-frequency representation (TFR) that represents the three-dimensional of the signal with respect to time and frequency in magnitude presentations. The manual lifting tasks is based on manual lifting of 5 kg and 10 kg load that performed by the right biceps brachii at lifting height of 75 cm and 140 cm. Four from ten healthy volunteers in fresh condition is selected into this comparison of subject performance tasks with their raw data collections. The raw data of EMG signals were then analyzed using MATLAB 2011 to obtain the voltage in time and frequency information. This study obtained the mean instantaneous RMS Voltage (Vrms(t)) to visualize the strength of the subjects produced during the manual lifting tasks. Results of this study evince the physical details of the subjects would able to effect the performance of the lifting. Higher lifting height and the number of contraction, the better performance of the subjects. It concluded that the application of spectrogram is able to providing the performance of the subjects by time-frequency information for EMG signals.
ISSN: 2180-1843 e-ISSN: 2289-8131 Vol. 8 No. 7 29
Performances Comparison of EMG Signal Analysis
for Manual Lifting using Spectrogram
T. N. S. T. Zawawi, A. R. Abdullah, E. F. Shair, S. M. Saleh
Faculty of Electrical Engineering, Universiti Teknikal Malaysia Melaka,
76100 Durian Tunggal, Melaka, Malaysia.
tgnorshuhadatgzawawi@gmail.com
AbstractElectromyography (EMG) signal is non-stationary
signal and highly complex time and frequency characteristics.
Fast-Fourier transform common technique in signal processing
involving EMG signal. However, this technique has a limitation
to provide the time-frequency information for EMG signals. This
paper presents the analysis of EMG signal of the variable lifting
height and mass of load between the four subjects selected in
manual lifting by using spectrogram. Spectrogram is one of the
time-frequency representation (TFR) that represents the three-
dimensional of the signal with respect to time and frequency in
magnitude presentations. The manual lifting tasks is based on
manual lifting of 5 kg and 10 kg load that performed by the right
biceps brachii at lifting height of 75 cm and 140 cm. Four from
ten healthy volunteers in fresh condition is selected into this
comparison of subject performance tasks with their raw data
collections. The raw data of EMG signals were then analyzed
using MATLAB 2011 to obtain the voltage in time and frequency
information. This study obtained the mean instantaneous RMS
Voltage (Vrms(t)) to visualize the strength of the subjects
produced during the manual lifting tasks. Results of this study
evince the physical details of the subjects would able to effect the
performance of the lifting. Higher lifting height and the number
of contraction, the better performance of the subjects. It
concluded that the application of spectrogram is able to
providing the performance of the subjects by time-frequency
information for EMG signals.
Index TermsElectromyography (EMG) Signal; Manual
Lifting; Spectrogram.
I. INTRODUCTION
Today, from the worldometers info, there are 7.3 billion
population in the world is estimated, with 3.1 billion number
of workers working in more than 55 major industrial sectors
[1]. In order providing job opportunities, a large number of
workplaces in the industrial sectors may lead to occupational
injuries if there is no awareness or concern regarding
occupational health and safety [1] and [2]. The National
Institute for Occupational Safety and Health (NIOSH)
recognized the growing problem of work-related back injuries
and published Work Practices Guide for Manual Lifting [3].
Manual lifting is commonly practiced by workers in
industrial workplace to move or transport good to a desired
place. In manual lifting, skeletal muscles perform a crucial
function to execute the task [4]. It is important to handle a
suitable load mass and lifting height to ensure the muscles
work in good experienced fatigue. Muscle performance is the
muscle’s endurance capability before the muscle experienced
fatigue [4].
Electromyography (EMG) signal have been widely used and
applied as a control signal in numerous man-machine interface
applications. It also been deployed in many clinical and
industrial applications [5]. The EMG is known as biomedical
signal that consist of electrical current that generated during
contraction and relaxation phase of muscles [6] and [7].
EMG signal is complicated and non-stationary signal with
highly complex time and frequency characteristics which is
controlled by nervous signal because it always responsible the
muscle activity [7] and [8]. This signal acquires noise and
distorts the signal while travelling through different tissues
during collection and recording process [6] and [9].
A lot of studies have been done based on EMG signal
investigation especially in extraction of EMG signal [10].
Previous researchers used fast Fourier transform (FFT) to
analyze the EMG signal, but it has the limitation to cater non-
stationary signals whose spectral characteristics change in
time [11]. It is no appropriate to use for non-stationary signal
as EMG. It only give the frequency information. Wavelet also
is common used in EMG, however it required more time to
analyze and required to find mother wavelet for each data. It is
also sensitive to noise while EMG produced noise while
travelling to each tissue [8], [9] and [12]. Besides that, to
extract the features, it involved high computational burden. By
artificial Intelligence (AI), it is higher complexity compared to
fast-Fourier transform and Wavelet [13]. Between all of this
common techniques in EMG signal processing, it is show that
Spectrogram with less complexity, better time and frequency
resolution and higher accuracy in order to get the information
from manual lifting activities.
In this research work, time frequency distributions (TFDs)
which are spectrogram employed to analyze the performance
of EMG signal for contraction of muscle activities. This
technique is one of the time-frequency representations (TFR)
that represent a three dimensional of the signal energy with
respect to time and frequency [14]. It is able to have superior
accuracy in challenging task colour modeling [15]. An
optimum frequency resolution is useful for extracting features
of any signals for further analysis included Electromyography
(EMG) signals [16]. From spectrogram, the overall
performance of the different mass of load, height and phase
angle would be known.
Journal of Telecommunication, Electronic and Computer Engineering
30 ISSN: 2180-1843 e-ISSN: 2289-8131 Vol. 8 No. 7
II. METHODOLOGY
A. Flow of data processing
Figure 1 shows the flow of the data collection and how the
data is processed to get the performance of EMG signal for the
contraction of muscle activity.
Figure 1: Flow of the data process
B. Subject Selection
Ten volunteers, five male and five female in healthy
conditions with no previous injuries participated in the
experimental work. G power application is an important
technical assumption for normality assumption for good
estimate the size of sample [17] and [18]. The number of
subjects was calculated using G power analysis software to
determine sufficient EMG raw signals for the analysis process.
Tens subjects selected based on the number of G Power
proposed between the age range of 22 to 25 years was selected
because this age range is commonly available in the
industries[1]. All subjects are right handed. The demographics
of the subjects is shown in [19]
C. Data Collection and Electrode Placement
The raw EMG data collection is measured by using surface
EMG (TeleMyo 2400T G2, Noraxon, USA) and MyoResearch
XP Master Software (Noraxon, USA). The electrode
positioning is at the right biceps. The details of the data
collection and electrodes placement is based on the previous
paper [20].
D. Manual Lifting Experiment
The subjects have to lift a 5 kg and 10kg load with a neutral
body twist (0º) symmetric lifting with the lifting height
75cm. The subject must repeat the lifting until achieve the
muscle fatigue for five times of repetition. Each lifting
produced EMG signals of contracted muscle, and each EMG
signal was divided into four phases as illustrated in the paper
[19] and [21]. At the Phase 1, the subject holds the load
(located on the floor). At Phase 2, subject lifts the load and put
it on the shelf, then Phase 3, subject arranges the load properly
on the shelf. Lastly, subject release the load at Phase 4.
III. EMG SIGNAL ANALYSIS TECHNIQUES
Spectrogram is used in this EMG signal processing because
it acquire to display the required information in EMG signal.
This is because FFT would not able to display the information
needed because of the limitation to non-stationary signals [22].
The details about spectrogram are already discussed in [19]
and [20]. The equation of the spectrogram is as followed:
2
2
( , ) ( ) ( )
xjf
S t f h w t e d

 


(1)
Instantaneous RMS voltage is extracted from spectrogram to
display the pattern of the signal of instantaneous value Vrms(t)
for the contraction of muscle activities behavioral. Vrms(t) can
be calculated using Equation (2) below [11]:
fmax
rms x
0
V (t)= S (t, f)dt
(2)
where Sx(t,f) is the time-frequency distribution and fmax is the
maximum frequency of interest.
IV. RESULTS AND DISCUSSION
The results of the EMG signal analysis are divided into
some parts to make it clearer presentation to know the
relationship between the different subjects and the
performance of the manual lifts.
A. EMG Signal of Manual Lifting
There are four parameters is considered in this experiment
for four subjects. Figure 2 until Figure 5 shows the example of
raw data for 75 cm and 140 cm lifting height with 5 kg and 10
kg load of mass. The maximum voltage for 5 kg 75 cm is
0.8x10-3 V and the minimum is 0.4x10-3 V with 36 times of
liftings. It is increasing peak voltage due to the increasing of
height with1.3x10-3 V maximum and minimum 0.6x10-3 V at
5 kg, 140 cm for 23 times of lifting.
Due to the Figure 4 (21 times lifting) and 5 (13 times
lifting), it is shown the raw data of manual lifting of 10 kg 75
START
Introduction (Literature review)
Data Collection- Manual lifting task
by TeleMyo 2400T G2, Noraxon,
USA and MyoResearch XP Master
Software (Noraxon, USA)
Linear Time-Frequency Distribution
Analysis
Time-Frequency Representation
Signal Parameter
END
Spectrogram
Performances Comparison of EMG Signal Analysis for Manual Lifting using Spectrogram
ISSN: 2180-1843 e-ISSN: 2289-8131 Vol. 8 No. 7 31
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
x 105
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2x 10-3
Sample (N)
Voltage (V)
RAW Signal (10kg, 140cm)
cm and 10 kg 140 cm. Their maximum for both parameters is
1.2x10-3 V and 1.4 x10-3 V while the minimum peak voltage is
1.1x10-3 V. Based on all the figures, it validates that the mass
of the load and the height of the lifting would able to affect the
peak value produced by the subjects.
Figure 6: RAW EMG signal for 5kg, 75cm
Figure 2: Raw EMG signal for 5 kg, 75 cm
Figure 3: Raw EMG signal for 5 kg, 140 cm
Figure 4: Raw EMG signal for 10kg, 75cm
Figure 5: Raw EMG signal for 10 kg, 140 cm
B. Time-Frequency Representations of EMG
Spectrogram is used in order to get the important
information in EMG signal processing analysis procedures.
The details about the spectrogram is discussed in the paper
[19] and [21]. Figure 6 show the time frequency representation
of spectrogram method in order to display the frequency, time
and amplitude for the EMG signals. Instantaneous RMS
Voltage (Vrms(t)) is extracted from spectrogram the displayed
the performance of the signals. It indicated the four phases
stated in the manual lifting experiments as shown in Figure 7.
Figure 6: Example of spectrogram analysis
Figure 7: Example of Instantaneous RMS Voltage (Vrms(t))
C. Performance Comparison of Manual Lifting Task
Mean is the parameter to tell the strength used of the muscle
for the EMG signal analysis [7]. Mean data are taken from the
mean Instantaneous Root Mean Square (RMS) Voltage
(Vrms(t)). The comparing is divided into four cases which are 5
kg 75 cm, 5 kg 140 cm, 10 kg, 75 cm and 10 kg 140 cm.
Subject 1 is sport man with the 174 cm height and thin,
Subject 2 is sport women with the 169 cm high thin, Subject 3
is sport man with the 166 cm high and fatter, last Subject 4 is
normal tough man 167cm fatter. In this result section, just two
phase is selected to be shown the example pattern of the
subject’s performance.
a. Manual Lifting of 5kg mass of load, 75cm lifting height
Figure 8(a), (b), represent the comparison from two from
four phase which is Phase 1 and Phase 2 to know the
performance of four subjects from ten involved in the manual
lifting tasks. In this tasks, it show that subject 1 and subject 2
more perform with the higher voltage and longer time to
fatigue (no of contractions) compared to subject 3 and 4. It is
increasing of mean (Vrms(t)) which is the strength when the
mass of load and lifting height in bigger.
In comparing this two selected phases, it show Phase 2
(Figure 8(b)) give higher strength compared to Phase 1(Figure
8(a)) because the subject have to lift the load that acquire more
strength compared to located the load at 5 kg and 75 cm.
0 0.5 1 1.5 2 2.5 3 3.5 4
x 105
-1.5
-1
-0.5
0
0.5
1
1.5x 10-3
Sample (N)
Voltage (V)
RAW Signal (5kg, 140cm)
0123456
x 105
-1
-0.5
0
0.5
1x 10-3
Sample (N)
Voltage (V)
RAW Signal (5kg, 75cm)
0 0.5 1 1.5 2 2.5 3 3.5
x 105
-1.5
-1
-0.5
0
0.5
1
1.5x 10-3
Sample (N)
Voltage (V)
RAW Signal (10kg, 75cm)
maximum
minimum
maximum
minimum
maximum
minimum
minimum
maximum
Time (ms)
Frequency (Hz)
Contour TFR
1000 2000 3000 4000 5000 6000 7000 8000 9000
10
20
30
40
50
60
70
80
90
100
1
2
3
4
5
6
7
8
x 10-10
01000 2000 3000 4000 5000 6000 7000 8000 9000 10000
0
1
2
3
4
5
6
7
8x 10-8
Time (ms)
Vrms(t)
Intanteneous RMS Voltage
Phase 1 Phase 2 Phase 3 Phase 4
Journal of Telecommunication, Electronic and Computer Engineering
32 ISSN: 2180-1843 e-ISSN: 2289-8131 Vol. 8 No. 7
(a)
(b)
Figure 8 (a) and (b) The performance of Mean (Vrms(t)) by the Subject 1,
Subject 2, Subject 3 and Subject 4
b. Manual Lifting of 5 kg mass of load, 75 cm lifting
height
The performance of four subjects involved is presented in
Figures 9(a) and (b). It show that Subject 1 and 2 still better
performance with higher strength and longer time to achieve
fatigue muscle. In this task, it is difference performance of
strength higher because at 140 cm lifting height, it acquire
more afford and harder for the subject with the lower high to
arrange the load onto the shelf, but for the taller subjects, it not
give much problem to them with 5 kg mass of load.
(a)
(b)
Figure 9 (a) and (b) The performance of Mean (Vrms(t)) of 5 kg 140 cm by
the Subject 1, Subject 2, Subject 3 and Subject 4
c. Manual Lifting of 10kg mass of load, 75cm lifting
height
Based on Figures 10(a) and (b), the subject’s performance
become different situation compared to the previous results.
From the subjects performances, it show that in heavier mass
of load, Subject 3 and 4 are higher performance compared to
Subject 1 and 2.
In this task, Subject 3 produced the second higher of
strength but longest time to achieve fatigue muscle. Subject 2
with highest strength but less the number of liftings. Subject 1
and 2 almost similar performance. The performance of the
subject almost similar the strength and although still
difference the time to fatigue muscle (number of lifting).
(a)
(b)
Figure 10 (a) and (b): The performance of Mean (Vrms(t)) of 10kg 75cm by
the Subject 1, Subject 2, Subject 3 and Subject 4
d. Manual Lifting of 10kg mass of load, 140cm lifting
height
In this task, it present the Subject 3 have used the highest
strength and longest time to fatigue compared to Subject 4,
Subject 2 and Subject 1. Subject 4 also quit higher strength
have used at the beginning of the manual lifting but drastically
decrease as same like the performance of Subject 1 and 2.
It is the same situation cross to the number of lifting can
produced by the subjects, the higher strength produced while
doing the lifting task the longer time of lifting to achieve
muscle fatigue.
0
0.00005
0.0001
0.00015
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35
SUBJECT 1 SUBJECT 2
SUBJECT 3 SUBJECT 4
0
0.00005
0.0001
0.00015
0.0002
0.00025
0.0003
0.00035
1357911 13 15 17 19 21 23 25 27 29 31 33 35
SUBJECT 1 SUBJECT 2
SUBJECT 3 SUBJECT 4
0
0.0001
0.0002
0.0003
0.0004
0.0005
1 2 3 4 5 6 7 8 9 1011 12 13 14 15 16 17 18 19 20 21 22 23 24
SUBJECT 1 SUBJECT 2 SUBJECT 3 SUBJECT 4
0
0.00005
0.0001
0.00015
1234567891011 12 13 14 15 16 17 18 19 20 21 22 23 24
SUBJECT 1 SUBJECT 2 SUBJECT 3 SUBJECT 4
0
0.0001
0.0002
0.0003
0.0004
0.0005
12345678910 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
SUBJECT 1 SUBJECT 2 SUBJECT 3 SUBJECT 4
0
0.0002
0.0004
0.0006
0.0008
0.001
0.0012
0.0014
1357911 13 15 17 19 21 23 25
SUBJECT 1 SUBJECT 2 SUBJECT 3 SUBJECT 4
Vrms
No. of lifting
No. of lifting
Vrms
No. of lifting
No. of lifting
Vrms
Vrms
No. of lifting
No. of lifting
Vrms
Vrms
Performances Comparison of EMG Signal Analysis for Manual Lifting using Spectrogram
ISSN: 2180-1843 e-ISSN: 2289-8131 Vol. 8 No. 7 33
(a)
(b)
Figure 11(a) and (b) The performance of Mean (Vrms(t)) of 10kg 140cm by
the Subject 1, Subject 2, Subject 3 and Subject
The comparison of all the performances for manual lifting
tasks is presented in Appendix A. Four subjects have selected
to be shown in the analysis process. The difference of task will
produce the different number of liftings. Each task is divided
into four phases as stated. Each will have the starting (1st
lifting) and ending (the end of lifting) mean Vrms(t) voltage
(strength). Either all phase, it would be focus on the highest
strength Phase 2 and Phase 3 which while travel the load and
arrange the load onto the shelf. All the subjects show by
increasing the mass of load and lifting height, the number of
lifting that would able to handle decreased but the strength
increase. At this phase it required more strength to handle the
load. V. CONCLUSION
Based on the analysis of EMG signals, this study concluded
that the application of spectrogram is able give the information
of the subject’s performance based on the variable lifting
height and mass of load from the determining of time-
frequency representation (TFR). From the summary of
Appendix A, it clearly show the relationship of the parameter
and physical condition of the subjects. All of this factors
would be affect the strength and time to fatigue. Subject 1 and
Subject 2 have the advantage to the higher lifting height, but
disadvantage to the higher mass of load. Furthermore, the
mean instantaneous RMS voltage (Vrms) which is strength
increased when the lifting height and mass of load is
increased. However, number of lifting would affected by the
strength produced by the subjects.
ACKNOWLEDGMENT
The authors wish to thank Universiti Teknologi Malaysia
(UTM) and Ministry of Higher Education Malaysia for
providing the funding and facilities to conduct this research
with Project No 05H51.
REFERENCES
[1] I. Halim, A. R. Omar, A. M. Saman, and I. Othman, “Assessment of
Muscle Fatigue Associated with Prolonged Standing in the Workplace,”
Saf. Health Work, vol. 3, no. 1, p. 31, 2012.
[2] I. Halim and A. R. Omar, “Development of prolonged standing strain
index to quantify risk levels of standing jobs,” Int. J. Occup. Saf. Ergon.,
vol. 18, no. 1, pp. 8596, 2012.
[3] R. Maiti and T. P. Bagchi, “Effect of different multipliers and their
interactions during manual lifting operations,” Int. J. Ind. Ergon., vol.
36, no. 11, pp. 9911004, 2006.
[4] M. S. Isa Halim, Rawaida, Kamat S. R, Rohana A., Adi Saptari,
“Analysis of Muscle Activity using Surface Electromyography for
Muscle Performance in Manual Lifting Task,” Appl. Mech. Mater., vol.
564, pp. 544649, 2014.
[5] A. Phinyomark, P. Phukpattaranont, and C. Limsakul, “Feature reduction
and selection for EMG signal classification,” Expert Syst. Appl., vol. 39,
no. 8, pp. 74207431, 2012.
[6] E. Gokgoz and A. Subasi, “Comparison of decision tree algorithms for
EMG signal classification using DWT,” Biomed. Signal Process.
Control, vol. 18, pp. 138144, 2015.
[7] S. D. Ruchika, “An Explanatory Study of the Parameters to Be Measured
From,” International Journal of Engineering And Computer Science,
vol. 2, no. 1, 2013.
[8] M. R. Canal, “Comparison of wavelet and short time Fourier transform
methods in the analysis of EMG signals,” J. Med. Syst., vol. 34, no. 1,
pp. 9194., 2010.
[9] M. B. I. Reaz, M. S. Hussain, and F. Mohd-Yasin, “Techniques of EMG
signal analysis: detection, processing, classification and applications,”
Biol. Proced. Online, vol. 8, no. 1, pp. 1135, 2006.
[10] D. C. R. E. Bekka, “Effect of the window length on the EMG spectral
estimation through the Blackman-Tukey method,” Signal Process. Its
Appl., vol. 2, pp. 1720, 2003.
[11] M. H. Abdullah, A.R.; Norddin, N.; Abidin, N.Q.Z.; Aman, A.; Jopri,
“Leakage current analysis on polymeric and non-polymeric insulating
materials using time-frequency distribution,” Power Energy (PECon),
2012 IEEE Int. Conf., no. December, pp. 25, 2012.
[12] A. Andreotti, A. Bracale, P. Caramia, and G. Carpinelli, “Adaptive
prony method for the calculation of power-quality indices in the
presence of nonstationary disturbance waveforms,” IEEE Trans. Power
Deliv., vol. 24, no. 2, pp. 874883, 2009.
[13] Ahsan, “Advances in Electromyogram Signal Classification to Improve
the Quality of Life for the Disabled and Aged People,” J. Comput. Sci.,
vol. 6, no. 7, pp. 706715, 2010.
[14] T. Srividya and A. M. Sankar, “Power Quality Analysis Using DSP
Techniques,” ITSI Trans. Electr. Electron. Eng., pp. 8086, 2013.
[15] and M. E. H. Joshi, Deepak, Bryson H. Nakamura, “High energy
spectrogram with integrated prior knowledge for EMG-based
locomotion classification,” Med. Eng. Phys., no. 5, pp. 518524, 2015.
[16] J. Kilby and K. Prasad, “Analysis of Surface Electromyography Signals
Using Discrete Fourier Transform Sliding Window Technique.,” Int. J.
Comput. Theory …, vol. 5, no. 2, pp. 321325, 2013.
[17] F. Faul, E. Erdfelder, A. Buchner, and A.-G. Lang, “Statistical power
analyses using G*Power 3.1: tests for correlation and regression
analyses.,” Behav. Res. Methods, vol. 41, no. 4, pp. 114960, 2009.
[18] F. Faul, E. Erdfelder, A.-G. Lang, and A. Buchner, “G*Power 3: a
flexible statistical power analysis program for the social, behavioral, and
biomedical sciences.,” Behav. Res. Methods, vol. 39, no. 2, pp. 17591,
2007.
[19] S. M. S. T.N.S.Tengku Zawawi, A.R.Abdullah, E.F. Shair, I. Halim,
“EMG Signal Analysis of Fatigue Muscle Activity in Manual Lifting,” J.
Electr. Syst., vol. 11, no. 3, pp. 319325, 2015.
[20] T.N.S.Tengku Zawawi, A.R.Abdullah, E. F. Shair, I. Halim, and R. O,
“Electromyography Signal Analysis Using Spectrogram,” IEEE Student
Conf. Res. Dev., no. December, pp. 1617, 2013.
[21] E. F. Shair, T. N. S. T. Zawawi, A. R. Abdullah, and N. H. Shamsudin,
“sEMG Signals Analysis Using Time-Frequency Distribution for
Symmetric and Asymmetric Lifting,” in 2015 International Symposium
on Technology Management and Emerging Technologies (ISTMET),
August 25 - 27, 2015, Langkawi, Kedah, Malaysia, 2015, pp. 233237.
0
0.0005
0.001
0.0015
0.002
12345678910 11 12 13 14 15 16 17 18
SUBJECT 1 SUBJECT 2 SUBJECT 3 SUBJECT 4
0
0.00005
0.0001
0.00015
0.0002
0.00025
12345678910 11 12 13 14 15 16 17 18
SUBJECT 1 SUBJECT 2 SUBJECT 3 SUBJECT 4
No. of lifting
No. of lifting
Vrms
Vrms
Journal of Telecommunication, Electronic and Computer Engineering
34 ISSN: 2180-1843 e-ISSN: 2289-8131 Vol. 8 No. 7
[22] N. Q. Z. Abidin, A. R. Abdullah, N. B. Norddin, and A. Aman, “Online
surface condition monitoring system using time-frequency analysis technique on high voltage insulators,” Power Eng. Optim. Conf.
(PEOCO), 2013 IEEE 7th Int., vol. 7, no. 11, pp. 513517, 2013.
APPENDIX
Appendix A shows the comparison of the performance of manual lifting tasks.
Parameters
Number of
liftings
Mean of (Vrms(t))
Phase 1
Phase 2
Phase 3
Phase 4
Start
End
Start
End
Start
End
Start
End
Subject 1
5 kg 75 cm
36
95.7
13.6
301.4
75.6
273.8
85.7
106.4
29.5
5 kg 140 cm
23
49.2
13.9
290.1
79.1
394.5
243.5
136.6
63.9
10 kg 75 cm
21
186.0
17.3
309.0
100.2
400.0
151.2
184.3
43.9
10 kg 140 cm
13
66.1
26.7
323.9
135.8
491.7
337.1
165.2
19.6
Subject 2
5 kg 75 cm
36
110.1
11.8
272.9
55.3
226.2
99.5
103.5
30.8
5 kg 140 cm
24
83.6
13.2
313.7
37.0
333.0
196.7
136.2
50.5
10 kg 75 cm
21
61.0
17.1
298.0
110.0
348.2
187.1
134.4
29.5
10 kg 140 cm
12
85.3
32.0
326.5
165.9
530.7
223.5
130.8
30.9
Subject 3
5 kg 75 cm
25
86.9
13.0
194.0
86.0
238.3
26.0
86.9
13.0
5 kg 140 cm
18
178.5
20.3
302.2
15.0
151.5
73.9
113.9
15.2
10 kg 75 cm
25
296.0
14.4
479.1
254.4
628.3
298.4
104.4
27.8
10 kg 140 cm
18
448.3
61.9
1183.3
619.2
1422.4
793.3
146.2
54.8
Subject 4
5 kg 75 cm
27
96.3
12.0
162.9
32.2
203.0
32.0
88.2
13.0
5 kg 140 cm
19
138.0
15.4
248.8
46.2
203.0
71.7
121.0
16.5
10 kg 75 cm
25
448.3
61.9
1183.3
619.2
1422.4
793.3
146.2
54.8
10 kg 140 cm
17
295.3
15.4
971.8
175.6
1280.8
125.9
207.7
29.6
... where X(f) is the EMG signal in the frequency domain and x(t) is the EMG signal in the time domain [2]. The STFT is a window function that provides the frequency content of chunks of a data set [3]. The FFT is performed on each individual chunk and computes a spectrum of the frequency vs. time data set [3]. ...
... The STFT is a window function that provides the frequency content of chunks of a data set [3]. The FFT is performed on each individual chunk and computes a spectrum of the frequency vs. time data set [3]. The STFT in discrete time is defined as: ...
... Where x[n] is the EMG signal in the time domain and w[n] is the window function [3]. The STFT is basically the Fourier Transform of the magnitude of the signal over time and frequency. ...
Conference Paper
Full-text available
The goal of this project is to create a more efficient method of analyzing EMG signals for use of clinical research and diagnosis of muscular disorders. An app will be developed that plots the spectrogram of EMG signals in real-time. Spectrograms give the time-frequency response of a signal, and with EMG signals being so complex, this method provides more features of the signal that engineers or medical professionals can extract and analyze. An EMG amplifier implemented on a PCB and surface electrodes will be used to collect EMG signals from the brachioradialis muscle. The EMG signals will then be sent to a PIC microprocessor where the information will be relayed via a Bluetooth modem to the Spectrogram app. The app will then calculate the short-time fast Fourier transform(STFT) of the signal and plot the spectrogram.
... For biomedical signals such as non-stationary EMG signal, spectrogram overcomes the limitation of time and frequency representation. More specifically, time and frequency domain present the signal in limited regular window size [16]. To obtain the time and frequency information simultaneously, TFD is preferred. ...
... Previous studies indicated lower window size affected the accuracy of frequency related information. However, a greater window size causes lower accurate in time related information [16]. In this analysis, Hanning window size of 128, 256 and 512 ms with 50% overlap are employed in order to determine the optimal time and frequency planes. ...
Article
Full-text available
This paper presents a study of the classification of myoelectric signal using spectrogram with different window sizes. The electromyography (EMG) signals of 40 hand movement types are collected from 10 subjects through NinaPro database. By employing spectrogram, the EMG signals are represented in time-frequency representation. Ten features are extracted from spectrogram for performance evaluation. In this study, two classifiers namely support vector machine (SVM) and linear discriminate analysis (LDA) are used to evaluate the performance of spectrogram features in the classification of EMG signals. To determine the best window size in spectrogram, three different Hanning window sizes are examined. The experimental results indicate that by applying spectrogram with optimize window size and LDA, the highest mean classification accuracy of 91.29% is obtained.
... However, TD and FD present in limited precision with regular window size [24]. To overcome this limitation, spectrogram is introduced to transform the EMG signal in timefrequency representation (TFR) [25]. Spectrogram is the most fundamental of the signal processing tool in noise and artifact reduction [26]. ...
Article
Full-text available
Electromyography (EMG) pattern recognition has recently drawn the attention of the researchers to its potential as an efficient manner in rehabilitation studies. In this paper, two time-frequency methods, discrete wavelet transform (DWT) and spectrogram are employed to obtain the time and frequency information from the EMG signal. Seventeen hand and wrist movements are recognized from the EMG signals acquired from ten intact subjects and eleven amputee subjects in NinaPro database. The root mean square (RMS) feature is extracted from each reconstructed DWT coefficient. On the other hand, the average energy of spectrogram at each frequency bin is extracted. The principal component analysis (PCA) preprocessing is applied to reduce the dimensionality of feature vectors. Four different classifiers namely Support Vector Machines (SVM), Decision Tree (DT), Linear Discriminate Analysis (LDA) and Naïve Bayes (NB) are used for classification. By applying SVM, DWT achieves the highest mean classification accuracy of 95% (intact subjects) and 71.3% (amputees). To validate our experimental results, the performance of DWT and spectrogram features are compared to other conventional methods. The obtained results obviously evince the superiority of DWT in EMG pattern recognition.
Article
Full-text available
The purpose of this paper is to improve the features of Health Screening Test System (HSTS) on Social Security Organization (SOCSO) program as physical evaluation for musculoskeletal disable workers (MSDs). SOCSO existing functional testing system are not suitable because of the evaluation was recorded manually peg board too high for Asian people. The occupational therapist whose involve in all the procedures is just doing the judgment in times to determine the capability of the patients. The functional capacity evaluation (FCE) technique is based on the functional range of motion evaluation that consist of positional tolerance respecting to time-motion testing on HSTS peg board and it is by referring to the original work movement. The main features of HSTS are able to measure speed, acceleration and evaluation of SOCSO’s patients for returning to work based on SOCSO’s requirement. In order to validate the accuracy of the proposed model, HSTS is used to evaluate the patient’s positional tolerance and then the result would be compare over the time frame which is Method Time Measurement (MTM) standard. This method is able to provide information and feedback for therapies as a status of patients. It is found that the proposed model is superior in getting the accurate time test for patient’s movements besides practicable and suitable for physical evaluation on MSD patients. © 2017 Institute of Advanced Engineering and Science. All rights reserved.
Article
Full-text available
In manufacturing industries, manual lifting is commonly practiced by workers in their routine to move or transport the objects to a desired place. Manual lifting with high repetition and loading on the arm will contribute the effects of soft tissues and muscle fatigue that will affect the performance of the worker to work with efficient. This paper presents the analysis of EMG signal from muscle activity to see the performance of muscle fatigue. Various researchers have proposed fast Fourier transforms (FFT) in analysing the EMG signal. However, this technique only gives spectral information but does not provide temporal information. Thus, the technique is not suitable for EMG analysis that consists of magnitude and frequency variation. To overcome the limitation, spectrogram is proposed to analyse the signal because it can represent the signal in jointly time-frequency representation (TFR). In fatigue muscle activities, ten volunteers in fresh condition and no previous of history injury are used as the subjects. Data is taken from right Biceps Branchii with lifting height of 140 cm and load mass of 5 kg. This research shows that the repeatability of manual lifting will contribute to the muscle fatigue for all the phases stated in this paper. This study concludes that phase 2 contribute highest effort by doing manual lifting task, compared to phase 1, 3 and 4, but all phases experienced the muscle fatigue.
Article
Full-text available
In industries, manual lifting is commonly practiced even though mechanized material handling equipment are provided. Manual lifting is used to transport or move products and goods to a desired place. Improper lifting techniquescontribute to muscle fatigue and low back pain that can lead to work efficiency and low productivity. The objective of this study were to analyze muscle activity in the left and right Erector Spinae, and left and right Biceps Brachii of five female subjects while performing manual lifting taskwithdifferent load mass, lifting height and twist angle. The muscle activitywere measured and analyzed using surface electromyography (sEMG).This study found that the right Biceps Brachii, right and left Erector Spinae experienced fatigue while performingasymmetric lifting (twist angle = 90°) at lifting height of 75 cm and 140 cm with load mass of 5 kg and 10 kg. Meanwhile, the left Biceps Brachii experienced fatigue when the lifting task was set at lifting height of 75 cm, load mass of 5 kg and twist angle of 90°.The load mass and lifting height has a significant influence to Mean Power Frequency (MPF) for left Biceps Brachii, left and right Erector Spinae. This study concluded that reducing the load mass can increase the muscles performance which can extend the transition-to-fatigue stage in the left and right Biceps Brachii and Erector Spinae.
Conference Paper
Full-text available
Electromyography (EMG) is known as complex bioelectricity signals that representing the contraction of the muscle in humanbody. The EMG signal offers useful information that can help to understand the human movement. Many techniques have been proposed by various researchers such as fast Fourier transforms (FFT). However, the technique only gives temporal information of the signal and does not suitable for EMG that consists of magnitude and frequency variation. In this paper,the analysis of EMG signal is presented using time-frequency distribution (TFD) which is spectrogram with different window size. Since the spectrogram represent the theEMG signal in time-frequency representation (TFR), it is very appropriate to analyze the signal. The EMG signals from Biceps muscle of two subjects are collected for body position of 0° and 90°. From the TFR, parameters of the signal such as instantaneous fundamental root mean square (RMS) voltage (Vrms) are estimated. To identify the suitable windows size, spectrogram with size window of 64, 256, 512 and 1024 is used to analyze the signal and the performance of the TFR are evaluated. The results show that spectrogram with window size of 512 gives optimal TFR of the EMG signals and suitable to characterize the signal.
Article
Full-text available
An optimum frequency resolution is useful for extracting features of any signals for further analysis. The purpose of this research was to determine the best frequency resolution of Surface Electromyography (sEMG) signals by using a sliding window with Discrete Fourier Transform to produce spectrogram plots. The process was carried out in two stages of investigations. The first investigation was to hold the sampling frequency constant at 8192 Hz and to use a new algorithm with sliding window sizes ranging from 16 to 512 samples through the signal. The results showed that the spectrogram that produced the best visual frequency resolution was with a window size of 64 samples. The calculated time period of sampling frequency of 8192 Hz with window size of 64 samples is hence 1/128 seconds. The second investigation was to use the time period of 1/128 seconds found in the investigation one. This time period of 1/128 second is held constant in order to determine window sizes that passed through different sampling frequencies which were set at 1024 Hz, 2048 Hz and 4096 Hz. Hence the calculated window sizes sample values are 8, 16 and 32 respectively. The spectrogram was plotted for each window sizes and it was found that a window size of 32 samples with the sampling frequency at 2048 Hz gave the best visual frequency resolution for the analysis of sEMG signals.
Conference Paper
Full-text available
This paper presents the performance of polymeric and non-polymeric insulation material in contaminated condition. The purpose of this study is to determine the accuracy of analysis technique on the performance these types of insulation against its tracking and erosion. Polypropylene and glass are selected as polymeric and non-polymeric material. Tracking and erosion test complying with BS EN 60587-2007 are conducted, then leakage current frequency components is used as a diagnostic tool for their surface condition monitoring and degradation severity. The analyses of leakage current are based on frequency domain and time-frequency representations are conducted. From the results, the surface condition of polymeric insulation material state can be classified accurately by using spectrogram compared to Fast Fourier Transform. It also proved that the performance of tracking and erosion of glass is better than polypropylene material.
Article
Electromyogram (EMG) signal representation is crucial in classification applications specific to locomotion and transitions. For a given signal, classification can be performed using discriminant functions or if-else rule sets, using learning algorithms derived from training examples. In the present work, a spectrogram based approach was developed to classify (EMG) signals for locomotion mode. Spectrograms for each muscle were calculated and summed to develop a histogram. If-else rules were used to classify test data based on a matching score. Prior knowledge of locomotion type reduced class space to exclusive locomotion modes. The EMG data were collected from seven leg muscles in a sample of able-bodied subjects while walking over ground (W), ascending stairs (SA) and the transition between (W-SA). Three muscles with least discriminating power were removed from the original data set to examine the effect on classification accuracy. Initial classification error was <20% across all modes, using leave one out cross validation. Use of prior knowledge reduced the average classification error to <11%. Removing three EMG channels decreased the classification accuracy by 10.8%, 24.3%, and 8.1% for W, W-SA, and SA respectively, and reduced computation time by 42.8%. This approach may be useful in the control of multi-mode assistive devices. Copyright © 2015 IPEM. Published by Elsevier Ltd. All rights reserved.
Article
Decision tree algorithms are extensively used in machine learning field to classify biomedical signals. De-noising and feature extraction methods are also utilized to get higher classification accuracy. The goal of this study is to find an effective machine learning method for classifying ElectroMyoGram (EMG) signals by applying de-noising, feature extraction and classifier. This study presents a framework for classification of EMG signals using multiscale principal component analysis (MSPCA) for de-noising, discrete wavelet transform (DWT) for feature extraction and decision tree algorithms for classification. The presented framework automatically classifies the EMG signals as myopathic, ALS or normal, using CART, C4.5 and random forest decision tree algorithms. Results are compared by using numerous performance measures such as sensitivity, specificity, accuracy, F-measure and area under ROC curve (AUC). Combination of DWT and random forest achieved the best performance using k-fold cross-validation with 96.67% total classification accuracy. These results demonstrate that the proposed approach has the capability for the classification of EMG signals with a good accuracy. In addition, the proposed framework can be used to support clinicians for diagnosis of neuromuscular disorders.
Article
G*Power (Erdfelder, Faul, & Buchner, 1996) was designed as a general stand-alone power analysis program for statistical tests commonly used in social and behavioral research. G*Power 3 is a major extension of, and improvement over, the previous versions. It runs on widely used computer platforms (i.e., Windows XP, Windows Vista, and Mac OS X 10.4) and covers many different statistical tests of the t, F, and chi2 test families. In addition, it includes power analyses for z tests and some exact tests. G*Power 3 provides improved effect size calculators and graphic options, supports both distribution-based and design-based input modes, and offers all types of power analyses in which users might be interested. Like its predecessors, G*Power 3 is free.