SlideShare a Scribd company logo
1 of 9
Download to read offline
International Journal of Electrical and Computer Engineering (IJECE)
Vol. 10, No. 6, December 2020, pp. 5694∼5702
ISSN: 2088-8708, DOI: 10.11591/ijece.v10i6.pp5694-5702 Ì 5694
Recognition of new gestures using myo armband for
myoelectric prosthetic applications
Jabbar Salman Hussain, Ahmed Al-Khazzar, Mithaq Nama Raheema
College of Engineering, University of Kerbala, Iraq
Article Info
Article history:
Received Feb 18, 2020
Revised May 5, 2020
Accepted May 19, 2020
Keywords:
Electromyography
Gestures
MLPNN
Multilayer perceptron neural
network
Myo armband
Myoelectric
Prosthetic
Recognition
ABSTRACT
Myoelectric prostheses are a viable solution for people with amputations. The chal-
lenge in implementing a usable myoelectric prosthesis lies in accurately recognizing
different hand gestures. The current myoelectric devices usually implement very few
hand gestures. In order to approximate a real hand functionality, a myoelectric prosthe-
sis should implement a large number of hand and finger gestures. However, increasing
number of gestures can lead to a decrease in recognition accuracy. In this work a
Myo armband device is used to recognize fourteen gestures (five build in gestures of
Myo armband in addition to nine new gestures). The data in this research is collected
from three body-able subjects for a period of 7 seconds per gesture. The proposed
method uses a pattern recognition technique based on Multi-Layer Perceptron Neural
Network (MLPNN). The results show an average accuracy of 90.5% in recognizing
the proposed fourteen gestures.
Copyright c 2020 Insitute of Advanced Engineeering and Science.
All rights reserved.
Corresponding Author:
Ahmed Al-Khazzar,
University of Kerbala, Kerbala, Iraq.
Email: ahmed.m.ahmed@uokerbala.edu.iq
1. INTRODUCTION
Upper-limb myoelectric prostheses are the next generation devices aimed at helping the amputees who
suffer from the loss of hand. Unlike the traditional hand prostheses which offer cosmetic or body powered lim-
ited functionality to the amputee, the myoelectric prostheses use electromyography (EMG) sensors attached
to the body muscles to control the prosthesis through a controller and motors that move the synthetic hand.
The mechanical aspect of the prosthetics has developed vastly in recent years and many prosthetic hands with
multiple movements and Degree of Freedom (DoF) are made available commercially. Examples of these pros-
thetic hands are Bebionic hand [1], iLimb hand [2], Vanderbilt hand [3], the UNB hand [4], the Yale hand [5],
SmartHand [6], the DLR/HIT hand [7], and the Keio hand [8].
Most of the currently developed myoelectric prostheses suffer from two problems. They either offer
a limited number of hand gestures or poor recognition rate for a larger number of gestures. Since the role of
a prostheses is to approximate the functionality of a real hand, there is a need for a larger number of gestures
to be recognized in a myoelectric prosthesis with high recognition rate. The current paper tries to address the
above problem.
Developing a controller for multiple hand movements that achieves acceptable recognition accuracy
is a challenge. The recognition rate and accuracy of EMG systems has been researched widely in the past [9].
According to [10], the recognition accuracy of such a system will depend mainly on a number of parameters:
EMG electrode location, the signal processing methods, feature extraction techniques, and the classification
algorithms.
Journal homepage: http://ijece.iaescore.com
Int J Elec & Comp Eng ISSN: 2088-8708 Ì 5695
According to [11], in surface EMG, the electrode position is vital for the detection of the signal and can
affect the quality of the signal. EMG data can be different when are obtained from different nearby locations
of the muscle. The use of multi-channel techniques can improve the quality of the signal with respect to the
position of the electrodes. Recently, in some myoelectric prostheses, the Myo armband device has been used
to collect EMG data. The Myo armband consists of eight electrodes which can mitigate the position issue.
Examining the previous work shows that most of the research in the past used between two to eight EMG
sensors in the forearm [12].
In regard to other parameters affecting the recognition accuracy, various feature extraction and clas-
sification algorithms were researched in the previous literature, as summarized in [13]. Recognizing multiple
gestures is a challenge as the recognition accuracy decreases with the increase in the number of gestures.
Table 1 summarizes examples of previous literature attempting to recognize several gestures [10, 14-25].
The proposed method in this work is to use multi-layer perceptron neural network (MLPNN) com-
bined with pre-processing techniques and feature extraction algorithms. In this research, an eight-channel Myo
armband EMG device is used to recognize fourteen hand gestures. These are nine gestures more than the five
built-in gestures the device can recognize. The novelty in this work lies in the recognition of fourteen gestures
with an advanced neural network with high recognition accuracy and improving upon the results reported in
the literature shown in Table 1. To the author’s knowledge, no myoelectric prosthesis is commercially available
today that provides multiple gesture recognition capability. This is due to various factors such as: signifi-
cant processing time of pattern recognition systems, the lack of reliability of these systems (due to electrode
positioning and fatigue), and the lack of intuitive controls [26-28].
Table 1. Recognition accuracy of previous literature with multiple gestures
Reference Features
No. of EMG
Channels
Classifier
No. of
Subjects
Time Length
(seconds)
No. of
Gestures
Recognition
Accuracy (%)
10 I-EMG, MAV, MSR, 6 PCA algorithm and 5 5 6 99.6
VAR, WL, SSI, SVM classifier 11 95.6
SSC, ZC, and WA 17 95.1
14 PSD 6 ANN 12 5 9 72.9
17 63.8
15 IRMS 4 ICA,IRAM,ANN 4 10 4 90.33
16 TD, ACCC, 4 KNN, LDA, SVM 6 0.064 9 91
and SPM
17 PSD-Av 5 PNM 4 1 10 80
18 HD-sEMG 8 MK-MMD 23 3-10 22 84.6
19 Third-order AR 2 Linear Bayesian 4 1 5 90-93.5
model Coefficients, Classifier 11 83.1-95.4
MAV and MAVR 16 78.8-90.3
20 CCA 4 KNN,LDA, and 8 5-10 8 82
LIBSNM
21 MAV 64 HD 5 2 9 78.21
22 MAV,ZC, 2 KNN 30 5 4 94
SSC, and WL
23 MV, WL 4 MLE 8 1-2 8 85.7
24 MAV, AR, 4 SVM,LDA, 18 1 8 89.3
and MNF and HMM
25 MAV,ARSSC 6 SVM 5 4 5 96
ZC,WL, and
RMS
13 RMS, STD, 8 MLPNN 3 5 5 99
MAX, and MIN
This work RMS 8 MLPNN 3 7 14 91-94
The advance of pattern recognition techniques can potentially help recognizing multiple DoF and
mapping them to myoelectric prostheses. However, recognizing gestures which include individual finger move-
ments using the EMG can be a challenging task. The reason is that the EMG signal amplitude variances are
small for finger movements compared to arm and wrist movements [14]. In addition, the muscles responsible
for controlling the movement of fingers are located in intermediate and deep layers of forearm [27]. In order to
recognize various finger movements, multiple EMG sensors are required to provide adequate data.
This paper is organized as follows. Section 2 presents the structure of the Myo armband device, and
the Multi-Layer Perceptron Neural Network (MLPNN). Section 3 presents data collection and the training of
the MLPNN. Results are discussed in Section 4. Finally the conclusion is presented.
Recognition of new gestures using myo armband for... (Jabbar Salman Hussain)
5696 Ì ISSN: 2088-8708
2. THEORY
2.1. Myo armband device
Myo armband is a wearable device that uses eight electrode sensors to measure the EMG signals of
forearm muscles. The device sends these data via Bluetooth to a computer. A special driver program is used to
analyze the signals and recognize hand gestures of the user. The recognized gestures can subsequently be used
by various applications. The Myo armband is also equipped with several other sensors, such as accelerometer,
gyroscope, and magnetometer which collect spatial data about the gesture and can facilitate gesture recognition
(these additional sensors were not utilized in this work). The device should be in direct contact with the user
skin, in order to operate and can work with several operating systems such as Windows, Mac, iOS, and Android.
It is powered by an ARM Cortex M4 processor and uses a rechargeable battery to operate [29-30].
2.2. Multi-layer perceptron neural network
Multi-layer perceptron neural network (MLPNN) is a feed-forward network which contains one or
more invisible layers called hidden layers. The learning process is performed using a supervised method where
the desired output must be known in advance to update the weights of the internal connections between the
layers. This update algorithm is called the backpropagation. In this algorithm the error between the desired
output and the actual output is calculated every time the input/output training data is presented. Connection
weights are continuously adjusted according to the calculated error until the error gradient reaches an appro-
priate small value, which indicates that the actual output is close to the target [31-34].
3. WORK
3.1. Data collection
In this work, data sets of fourteen gestures were collected from three body-able subjects using Myo
armband device, as shown in Figure 1. The first five gestures (1 to 5) were selected based on the standard
Myo Armband gestures. The additional nine gestures were selected based on the natural movement of the hand
and fingers and some local cultural gestures. Each of the fourteen gestures were repeated twice, and the EMG
signal was recorded for a period of seven seconds per gesture with a frequency of 200Hz. Table 2 represents
the hand gesture categories and their description.
Table 2. Hand gestures categories and the description
Modes Gestures Description
Relax Rest The relax state
Wrist Movements Wave right Wave hand to the right
Wave Left Wave hand to the left
Individual Finger Movements One Finger Raise the index finger up
Thumb Raise the thumb up
Multi-Fingers Movement Spread Fingers Spread all fingers
Fist Fist state
Two Fingers Raise the index and middle fingers up
Three Fingers Raise the index, middle and ring fingers up
Four Fingers Raise all fingers except the thumb up
Grab Grab gesture (a glass or bottle)
Pinch Pinch the fingers
Shake no Thumb Shake state while holding thumb inside
Shake Hand Shake state with thumb up
Each gesture’s data was stored as a sequence with a length varying between 1398 to 1408 EMG values
(elements). A remove-silence process was applied to the data followed by removing the extra elements to create
equal data sets. Furthermore, the elements in the data sets were overlapped to maximize number of samples.
This yielded a data set with a dimension of 50x117 per gesture per EMG channel.
Four features, namely Root Mean Square (RMS), Standard Deviation (STD), minimum (Min) and
maximum (Max) values, were extracted from each gesture sequence. This process has obtained a (4x117)
matrix of features per each EMG channel. For 8 channels, the result was a (32x117) features’ matrix per
gesture. 80% of the obtained data were used for training (input vector), 10% were used for validation and the
remaining 10% were used for testing the neural network.
Int J Elec & Comp Eng, Vol. 10, No. 6, December 2020 : 5694 – 5702
Int J Elec & Comp Eng ISSN: 2088-8708 Ì 5697
Figure 1. Representation of the fourteen gestures used in the work
3.2. Training of MLPNN
For the pattern recognition, a classifier was designed using the MLPNN and based on time domain
features of RMS, STD, Min and Max. The structure of the MLPNN was configured with one hidden layer to
reduce processing time in the training process. The number of hidden neurons was selected based on trial and
error and from experience to achieve the lowest RMS error possible. In addition, the aim was to minimize the
number of training epochs in the network.
Figure 2 shows the neural structure that was used for training. The classifier consisted of 32 neurons
in the input layer, 30 hidden processors with tangent function, and 14 output neurons with sigmoidal function
to classify 14 determined gestures. Gradient descent back-propagation was used as a training algorithm with
an adaptive learning rate to minimize the risks of local minima errors.
The training must be terminated if any of the stopping criteria is obtained to prevent the over fitting
problem. These criteria were set to 1000 training iterations and an RMS error of 0.001. Figure 3 shows the
algorithms used for training and testing of the MLPNN classifier.
Recognition of new gestures using myo armband for... (Jabbar Salman Hussain)
5698 Ì ISSN: 2088-8708
Figure 2. Structure of neural classifier
Begin
Prepare the Myo
Armband sensor
Read the EMG data
(training samples)
Training MLPNN
No
Yes
Stopping Criteria
 is met?
Save the trained
MLPNN
End
Begin
Read the new EMG
data (test samples)
Perform gesture
recognition using the
trained MLPNN
classifier
No
Yes
Test another 
gesture?
End
Pre-Processing
Segmentation
Overlapping
Feature Extraction
(a) (b)
Pre-Processing
Segmentation
Overlapping
Feature Extraction
Output the
recognition decision
Figure 3. Algorithms used for (a) the training, and (b) the testing of MLPNN classifier
4. RESULTS
Figure 4 shows the performance of the network at training, validation, and test processes. It can be
observed that the network was successfully trained with a validation performance of 0.001541. Figure 5 shows
the error gradient for all training epochs with the RMS value of 0.00035374 at epoch 101. Number of validation
checks was set to six. This validation check was achieved at epoch 101 as shown in the figure.
Int J Elec & Comp Eng, Vol. 10, No. 6, December 2020 : 5694 – 5702
Int J Elec & Comp Eng ISSN: 2088-8708 Ì 5699
Figure 4. Performance of the neural network at training, validation, and test processes
Figure 5. Results of training MLPNN: the x-axis shows the number of epochs, the y-axis is the gradient of
error at the top and the validation check value at bottom subfigure
Figure 6 shows the real performance of the MLPNN at the end of the training and testing processes.
The gesture number and the class number are represented at X-axis and Y-axis respectively. The training results
provide a perfect classification with high accuracy for all of the gestures. The class number of the gesture at
any testing process is defined by the largest value from all neurons at the output layer of MLPNN. For binary
patterns, the output of the categorized gesture must have a value of one, and the output of the other gestures
must have a value of zero. Table 3 shows the numerical representation of gesture classes.
Recognition of new gestures using myo armband for... (Jabbar Salman Hussain)
5700 Ì ISSN: 2088-8708
Figure 6. Real performance of MLPNN at (a) training process, (b) testing process
The segmentation and overlapping methods that were used in this work as part of the data pre-
processing improved upon the training efficiency by increasing the number of training samples. The accuracy
of the trained MLPNN is evaluated by calculating the number of correctly recognized classes in the test pro-
cess. Thus the recognition rate is calculated by:
RecognitionRate = (CorrectlyClassifiedSamples/AllTestSamples) ∗ 100 (1)
Table 4 shows the recognition rates results for the three body able subjects. The recognition accuracy
was up to 94%. Comparing these results to the previous EMG based recognition systems utilizing neural
networks shows that a better recognition accuracy was achieved in this work with respect to number of gestures
that can be recognized as shown in Table 5. To the authors knowledge, there were no work available using the
Myo armband for recognition of multiple gestures for comparison purposes.
The previous work of the same authors reported 99% of accuracy using similar neural network and 5
gestures. Hence adding 9 new gestures decreased the accuracy of the system to 90.5%. The decreased accuracy
can be attributed to an increase in the number of gestures that are differentiated based on finger movements.
The EMG signal amplitude variances are small for finger movements compared to arm and wrist movements
which results in similar extracted features and less accurate recognition.
Table 3. Assigned class numbers to the gestures
Gestures Rest
Spread
fingers
Fist
Wave
Right
Wave
Left
One
Finger
Two
Finger
Three
Finger
Four
Finger
Thumb Grab Pinch
Shake no
Thumb
Shake
Hand
Real O/P 1 2 3 4 5 6 7 8 9 10 11 12 13 14
Table 4. Recognition rates of the three subjects
Subject No.
1st Set
Recognition Rate %
2nd Set
Recognition Rate %
Recognition
Average %
1 89 89 89
2 89 91 90
3 91 94 92.5
Table 5. Comparison of results with other ANN classifiers in the literature
Reference
No. of EMG
Channels
No. of
Subjects
Time Length
(seconds)
No. of
Gestures
Recognition
Accuracy (%)
14 6 12 5
9 72.9
17 63.8
15 I4 4 10 4 90.33
13 8 3 5 5 99
This work 8 3 7 14 90.5
Int J Elec & Comp Eng, Vol. 10, No. 6, December 2020 : 5694 – 5702
Int J Elec & Comp Eng ISSN: 2088-8708 Ì 5701
5. CONCLUSION
This research proposed using an MLPNN to classify and recognize additional gestures from the Myo
armband device. The proposed system works by measuring the raw EMG signals from the forearm and esti-
mating the muscles activity by extracting the effective time domain features such as RMS, STD, Min and Max.
Then the system trains the MLPNN to classify different finger and wrist gestures. The main contribution of this
work is the accurate recognition of fourteen gestures using the Myo armband device which can only recognize
five gestures. The proposed pre-processing technique based on time domain feature extraction improved the
training process and was effective for the recognition of gestures.
The results showed an average recognition accuracy of 90.5% for all the test data. Based on these
results, the proposed method can be used as a reliable classifier for different hand gestures in applications
where recognizing as many gestures as possible is important such as multi-gesture prosthetic hands. The
implemented system can also contribute to the advancing of the intelligent interaction between the human and
instruments such as myoelectric prosthetic hands, machine control and Human Computer Interface (HCI).
Future work can design and implement a manageable system for recognition of hand gestures using
hardware implementation techniques such as Field Programmable Gate Array (FPGA). The error in recognition
could be due to convergent gestures. The trained MLPNN may mis-classify these gestures as a result of similar
features calculated from comparable raw EMG signals due to the contiguity of their associative muscles. The
mis-recognized gestures vary depending on the subject under test, since each subject could perform the same
gesture slightly different each time. An extension to this work can research methods to deal with these incorrect
recognitions in order to improve the accuracy of the system. Finally, An MPLNN based recognition algorithm
trained by back-propagation could suffer from local minima or high computational cost. Future work needs
to examine other intelligent algorithms such as Adaptive Neuro Fuzzy Inference System (ANFIS) for gesture
recognition.
ACKNOWLEDGEMENT
This work is carried out by Intelligent Prosthetics Research Group (IPRG) at College of Engineering,
University of Kerbala in 2019.
REFERENCES
[1] Bebionic Hand, Ottobock, [Online] Available: https://www.ottobockus.com/prosthetics/upper-limb-
prosthetics/solution-overview/bebionic-hand.
[2] I-Limb Ultra, Ossur, [Online] Available: https://www.ossur.com/en-us/prosthetics/arms/i-limb-ultra.
[3] S. A. Dalley, et al. “Design of a multifunctional anthropomorphic prosthetic hand with extrinsic actua-
tion,” IEEE/ASME Transactions on Mechatronics, vol. 14, no. 6, pp. 699-706, 2009.
[4] Y. Losier, et al., “An overview of the UNB hand system,” Proc. of the Myoelectric Controls/Powered
Prosthetics Symp., 2011.
[5] J. T. Belter, and A. M. Dollar, “Novel differential mechanism enabling two DoF from a single actuator:
Application to a prosthetic hand,” IEEE 13th Int. Conf. on Rehabilitation Robotics, pp. 1-6, 2013.
[6] C. Cipriani, M. Controzzi, and M.C Carrozza, “The SmartHand transradial prosthesis,” Journal of neuro-
engineering and rehabilitation, vol. 8, no. 1, 2011.
[7] H. Liu, et al., “Multisensory five-finger dexterous hand: The DLR/HIT Hand II,” IEEE/RSJ international
conference on intelligent robots and systems, pp. 3692-3697, 2008.
[8] Y. Kamikawa, and T. Maeno, “Underactuated five-finger prosthetic hand inspired by grasping force dis-
tribution of humans,” IEEE/RSJ Int. Conf. on Intelligent Robots and Systems, pp. 717-722, 2008.
[9] N. E. Krausz, and R. A. Rorrer, “Design and fabrication of a six degree-of-freedom open source hand,”
IEEE Transactions on Neural Systems and Rehabilitation Engineering, vol. 24, no. 5, pp. 562-572, 2015.
[10] Z. Chen, et al., “Hand gestures recognition from multi-channel forearm EMG signals,” International
Conference on Cognitive Systems and Signal Processing, pp. 119-125, 2016.
[11] L. Mesin, R. Merletti, and A. Rainoldi, “Surface EMG: the issue of electrode location,” Journal of Elec-
tromyography and Kinesiology, vol. 19, no. 5, 2009.
[12] X. Zhang, et al., “A framework for hand gesture recognition based on accelerometer and EMG sensors,”
IEEE Transactions on Systems, Man, and Cybernetics-Part A: Systems and Humans, vol. 41, no. 6,
pp. 1064-1076, 2011.
Recognition of new gestures using myo armband for... (Jabbar Salman Hussain)
5702 Ì ISSN: 2088-8708
[13] A.M. AlKhazzar and M.N. Raheema, “EMG Signal Classification Using Radial Basis Function Neural
Network,” Third Scientific Conference of Electrical Engineering (SCEE), pp. 180-185, 2018.
[14] T. Jiralerspong, et al., “Experimental study of real-time classification of 17 voluntary movements for
multi-degree myoelectric prosthetic hand,” Applied Sciences, vol. 7, no. 1, 2017.
[15] G. R. Naik, D. K. Kumar, and S. P. Arjunan, “Multi modal gesture identification for HCI using surface
EMG,” Proc. of the 12th Int. conf. on entertainment and media in the ubiquitous era, pp. 90-94, 2008.
[16] X. Chen, and Z.J. Wang, “Pattern recognition of number gestures based on a wireless surface EMG
system,” Biomedical Signal Processing and Control, vol. 8, no. 2, pp. 184-192, 2013.
[17] M.C. Castro, S.P. Arjunan, and D.K. Kumar, “Selection of suitable hand gestures for reliable myoelectric
human computer interface,” Biomedical engineering online, vol. 14, no. 1, 2015
[18] Y. Du, et al., “Surface EMG-based inter-session gesture recognition enhanced by deep domain adapta-
tion,” Sensors, vol. 17, no. 3, 2017.
[19] X. Chen, et al., “Hand gesture recognition research based on surface EMG sensors and 2D-
accelerometers,” 11th IEEE International Symposium on Wearable Computers, pp. 11-14, 2007.
[20] N. K. Rami, “Correlation Analysis of Electromyogram Signals for Multiuser Myoelectric Interfaces,”
IEEE tran. on neural systems and rehabilitation eng., vol. 22, no. 4, pp. 745-755, 2014.
[21] A. Moin, et al., “Analysis of Contraction Effort Level in EMG-Based Gesture Recognition Using Hyper-
dimensional Computing,” IEEE biomedical circuits and systems conf., 2019.
[22] W.T. Shi, et al., “A bionic hand controlled by hand gesture recognition based on surface EMG signals: A
preliminary study,” Biocybernetics and Biomedical Engineering, vol. 38, no. 1, pp. 126-135, 2018.
[23] K. Rhee, and H.C. Shin, “Electromyogram-based hand gesture recognition robust to various arm pos-
tures,” International Journal of Distributed Sensor Networks, vol. 14, no. 7, 2018.
[24] W. Geng, Du, Y., Jin, W., Wei, W., Hu, Y., and Li, J, “Gesture recognition by instantaneous surface EMG
images,” Scientific reports, vol. 6, 2016.
[25] X. Luo, et al., “Synergistic Myoelectrical Activities of Forearm Muscles Improving Robust Recognition
of Multi-Fingered Gestures,” Sensors, vol. 19, no. 3, pp. 610, 2019.
[26] A.D. Roche, et al., “Prosthetic myoelectric control strategies: a clinical perspective,” Current Surgery
Reports, vol. 2, no. 3, pp. 44, 2014.
[27] R. L. Drake, A. W. Vogl, and A. W. Mitchell, “Gray’s anatomy for students 2nd ed.,” Churchill Living-
stone, 2009.
[28] A.H. Al-Timemy, et al., “Improving the performance against force variation of EMG controlled multi-
functional upper-limb prostheses for transradial amputees,” IEEE Transactions on Neural Systems and
Rehabilitation Engineering, vol. 24, no. 6, pp. 650-661, 2015.
[29] S. Rawat, S. Vats and P. Kumar, ”Evaluating and exploring the MYO ARMBAND,” Int. Conf. System
Modeling and Advancement in Research Trends (SMART), pp. 115-120, 2016.
[30] “Myo Gesture Control Armband tech specs,” [Online] Available: https://support.getmyo.com/ hc/en-
us/articles/202648103-Myo-Gesture-Control-Armband-tech-specs.
[31] N. F. Ebecken, “An overview on the use of neural networks for data mining tasks,” Journal of the Brazilian
Neural Network Society, vol. 9, no. 3, pp. 202-212, 2011.
[32] G. Panchal, et al., “Behaviour analysis of multilayer perceptrons with multiple hidden neurons and hidden
layers,” Int. Journal of Computer Theory and Eng., vol. 3, no. 2, pp. 332-337, 2011.
[33] M.A. Hossain, et al., “Implementation Of Back-Propagation Neural Network For Isolated Bangla Speech
Recognition,” Int. Journal of Information Sciences and Techniques (IJIST), vol. 3, no. 4, 2013.
[34] M.C. Popescu, et al., “Multilayer perceptron and neural networks,” Wseas Transactions on Circuits and
Systems, vol. 8, no. 7, pp. 579-588, 2009.
Int J Elec & Comp Eng, Vol. 10, No. 6, December 2020 : 5694 – 5702

More Related Content

What's hot

Brain computer interface based
Brain computer interface basedBrain computer interface based
Brain computer interface based
Mohammed Saeed
 
NEUROPROSTHETICS
NEUROPROSTHETICSNEUROPROSTHETICS
NEUROPROSTHETICS
Haris Tyeb
 
Iaetsd recognition of emg based hand gestures
Iaetsd recognition of emg based hand gesturesIaetsd recognition of emg based hand gestures
Iaetsd recognition of emg based hand gestures
Iaetsd Iaetsd
 
Analysis of EEG data Using ICA and Algorithm Development for Energy Comparison
Analysis of EEG data Using ICA and Algorithm Development for Energy ComparisonAnalysis of EEG data Using ICA and Algorithm Development for Energy Comparison
Analysis of EEG data Using ICA and Algorithm Development for Energy Comparison
ijsrd.com
 
Final Draft of Research Paper 1
Final Draft of Research Paper 1Final Draft of Research Paper 1
Final Draft of Research Paper 1
Trevor Davis
 
FUSION OF GAIT AND FINGERPRINT FOR USER AUTHENTICATION ON MOBILE DEVICES
FUSION OF GAIT AND FINGERPRINT FOR USER AUTHENTICATION ON MOBILE DEVICESFUSION OF GAIT AND FINGERPRINT FOR USER AUTHENTICATION ON MOBILE DEVICES
FUSION OF GAIT AND FINGERPRINT FOR USER AUTHENTICATION ON MOBILE DEVICES
vasim hasina
 
Implementation and Evaluation of Signal Processing Techniques for EEG based B...
Implementation and Evaluation of Signal Processing Techniques for EEG based B...Implementation and Evaluation of Signal Processing Techniques for EEG based B...
Implementation and Evaluation of Signal Processing Techniques for EEG based B...
Damian Quinn
 

What's hot (20)

IRJET- A Survey on Control of Mechanical ARM based on Hand Gesture Recognitio...
IRJET- A Survey on Control of Mechanical ARM based on Hand Gesture Recognitio...IRJET- A Survey on Control of Mechanical ARM based on Hand Gesture Recognitio...
IRJET- A Survey on Control of Mechanical ARM based on Hand Gesture Recognitio...
 
Brain computer interface based
Brain computer interface basedBrain computer interface based
Brain computer interface based
 
NEUROPROSTHETICS
NEUROPROSTHETICSNEUROPROSTHETICS
NEUROPROSTHETICS
 
Iaetsd recognition of emg based hand gestures
Iaetsd recognition of emg based hand gesturesIaetsd recognition of emg based hand gestures
Iaetsd recognition of emg based hand gestures
 
Robot Motion Control Using the Emotiv EPOC EEG System
Robot Motion Control Using the Emotiv EPOC EEG SystemRobot Motion Control Using the Emotiv EPOC EEG System
Robot Motion Control Using the Emotiv EPOC EEG System
 
IRJET- Automation of a Prosthetic Limb using Shared Control of Brain Mach...
IRJET-  	  Automation of a Prosthetic Limb using Shared Control of Brain Mach...IRJET-  	  Automation of a Prosthetic Limb using Shared Control of Brain Mach...
IRJET- Automation of a Prosthetic Limb using Shared Control of Brain Mach...
 
Analysis of EEG data Using ICA and Algorithm Development for Energy Comparison
Analysis of EEG data Using ICA and Algorithm Development for Energy ComparisonAnalysis of EEG data Using ICA and Algorithm Development for Energy Comparison
Analysis of EEG data Using ICA and Algorithm Development for Energy Comparison
 
Bionic arm using muscle sensor v3
Bionic arm using muscle sensor v3Bionic arm using muscle sensor v3
Bionic arm using muscle sensor v3
 
Ijetcas14 393
Ijetcas14 393Ijetcas14 393
Ijetcas14 393
 
Final Draft of Research Paper 1
Final Draft of Research Paper 1Final Draft of Research Paper 1
Final Draft of Research Paper 1
 
Dsp lab report- Analysis and classification of EMG signal using MATLAB.
Dsp lab report- Analysis and classification of EMG signal using MATLAB.Dsp lab report- Analysis and classification of EMG signal using MATLAB.
Dsp lab report- Analysis and classification of EMG signal using MATLAB.
 
FUSION OF GAIT AND FINGERPRINT FOR USER AUTHENTICATION ON MOBILE DEVICES
FUSION OF GAIT AND FINGERPRINT FOR USER AUTHENTICATION ON MOBILE DEVICESFUSION OF GAIT AND FINGERPRINT FOR USER AUTHENTICATION ON MOBILE DEVICES
FUSION OF GAIT AND FINGERPRINT FOR USER AUTHENTICATION ON MOBILE DEVICES
 
Final thesis presentation on bci
Final thesis presentation on bciFinal thesis presentation on bci
Final thesis presentation on bci
 
K44095156
K44095156K44095156
K44095156
 
Wavelet-based EEG processing for computer-aided seizure detection and epileps...
Wavelet-based EEG processing for computer-aided seizure detection and epileps...Wavelet-based EEG processing for computer-aided seizure detection and epileps...
Wavelet-based EEG processing for computer-aided seizure detection and epileps...
 
Bio-medical (EMG) Signal Analysis and Feature Extraction Using Wavelet Transform
Bio-medical (EMG) Signal Analysis and Feature Extraction Using Wavelet TransformBio-medical (EMG) Signal Analysis and Feature Extraction Using Wavelet Transform
Bio-medical (EMG) Signal Analysis and Feature Extraction Using Wavelet Transform
 
Design and Implementation of Brain Computer Interface for Wheelchair control
Design and Implementation of Brain Computer Interface for Wheelchair controlDesign and Implementation of Brain Computer Interface for Wheelchair control
Design and Implementation of Brain Computer Interface for Wheelchair control
 
Implementation and Evaluation of Signal Processing Techniques for EEG based B...
Implementation and Evaluation of Signal Processing Techniques for EEG based B...Implementation and Evaluation of Signal Processing Techniques for EEG based B...
Implementation and Evaluation of Signal Processing Techniques for EEG based B...
 
IRJET= Air Writing: Gesture Recognition using Ultrasound Sensors and Grid-Eye...
IRJET= Air Writing: Gesture Recognition using Ultrasound Sensors and Grid-Eye...IRJET= Air Writing: Gesture Recognition using Ultrasound Sensors and Grid-Eye...
IRJET= Air Writing: Gesture Recognition using Ultrasound Sensors and Grid-Eye...
 
Classification of EEG Signals for Brain-Computer Interface
Classification of EEG Signals for Brain-Computer InterfaceClassification of EEG Signals for Brain-Computer Interface
Classification of EEG Signals for Brain-Computer Interface
 

Similar to Recognition of new gestures using myo armband for myoelectric prosthetic applications

Comparative analysis of machine learning algorithms on myoelectric signal fro...
Comparative analysis of machine learning algorithms on myoelectric signal fro...Comparative analysis of machine learning algorithms on myoelectric signal fro...
Comparative analysis of machine learning algorithms on myoelectric signal fro...
IAESIJAI
 

Similar to Recognition of new gestures using myo armband for myoelectric prosthetic applications (20)

Using deep neural networks in classifying electromyography signals for hand g...
Using deep neural networks in classifying electromyography signals for hand g...Using deep neural networks in classifying electromyography signals for hand g...
Using deep neural networks in classifying electromyography signals for hand g...
 
Comparative analysis of machine learning algorithms on myoelectric signal fro...
Comparative analysis of machine learning algorithms on myoelectric signal fro...Comparative analysis of machine learning algorithms on myoelectric signal fro...
Comparative analysis of machine learning algorithms on myoelectric signal fro...
 
Short-term hand gestures recognition based on electromyography signals
Short-term hand gestures recognition based on electromyography signalsShort-term hand gestures recognition based on electromyography signals
Short-term hand gestures recognition based on electromyography signals
 
Prosthetic hand using Artificial Neural Network
Prosthetic hand using Artificial Neural NetworkProsthetic hand using Artificial Neural Network
Prosthetic hand using Artificial Neural Network
 
IRJET-V9I114.pdfA Review Paper on Economical Bionic Arm with Predefined Grasp...
IRJET-V9I114.pdfA Review Paper on Economical Bionic Arm with Predefined Grasp...IRJET-V9I114.pdfA Review Paper on Economical Bionic Arm with Predefined Grasp...
IRJET-V9I114.pdfA Review Paper on Economical Bionic Arm with Predefined Grasp...
 
Motor Imagery based Brain Computer Interface for Windows Operating System
Motor Imagery based Brain Computer Interface for Windows Operating SystemMotor Imagery based Brain Computer Interface for Windows Operating System
Motor Imagery based Brain Computer Interface for Windows Operating System
 
Prosthetic Arm for Amputees
Prosthetic Arm for AmputeesProsthetic Arm for Amputees
Prosthetic Arm for Amputees
 
WAVELET DECOMPOSITION METHOD BASED AUTOMATED DIAGNOSIS OF MUSCLE DISEASES
WAVELET DECOMPOSITION METHOD BASED AUTOMATED DIAGNOSIS OF MUSCLE DISEASESWAVELET DECOMPOSITION METHOD BASED AUTOMATED DIAGNOSIS OF MUSCLE DISEASES
WAVELET DECOMPOSITION METHOD BASED AUTOMATED DIAGNOSIS OF MUSCLE DISEASES
 
40120130406015
4012013040601540120130406015
40120130406015
 
Design a System for Hand Gesture Recognition with Neural Network
Design a System for Hand Gesture Recognition with Neural NetworkDesign a System for Hand Gesture Recognition with Neural Network
Design a System for Hand Gesture Recognition with Neural Network
 
EMG electromayogram
EMG electromayogramEMG electromayogram
EMG electromayogram
 
Electroencephalography-based brain-computer interface using neural networks
Electroencephalography-based brain-computer interface using neural networksElectroencephalography-based brain-computer interface using neural networks
Electroencephalography-based brain-computer interface using neural networks
 
Medical-Robotics
Medical-RoboticsMedical-Robotics
Medical-Robotics
 
RECOGNITION SYSTEM USING MYO ARMBAND FOR HAND GESTURES - SURVEY
RECOGNITION SYSTEM USING MYO ARMBAND FOR HAND GESTURES - SURVEYRECOGNITION SYSTEM USING MYO ARMBAND FOR HAND GESTURES - SURVEY
RECOGNITION SYSTEM USING MYO ARMBAND FOR HAND GESTURES - SURVEY
 
Hand motion pattern recognition analysis of forearm muscle using MMG signals
Hand motion pattern recognition analysis of forearm muscle using MMG signalsHand motion pattern recognition analysis of forearm muscle using MMG signals
Hand motion pattern recognition analysis of forearm muscle using MMG signals
 
Modelling and Control of a Robotic Arm Using Artificial Neural Network
Modelling and Control of a Robotic Arm Using Artificial Neural NetworkModelling and Control of a Robotic Arm Using Artificial Neural Network
Modelling and Control of a Robotic Arm Using Artificial Neural Network
 
Brain-computer interface of focus and motor imagery using wavelet and recurre...
Brain-computer interface of focus and motor imagery using wavelet and recurre...Brain-computer interface of focus and motor imagery using wavelet and recurre...
Brain-computer interface of focus and motor imagery using wavelet and recurre...
 
Multilayer extreme learning machine for hand movement prediction based on ele...
Multilayer extreme learning machine for hand movement prediction based on ele...Multilayer extreme learning machine for hand movement prediction based on ele...
Multilayer extreme learning machine for hand movement prediction based on ele...
 
A Review of Lie Detection Techniques.pdf
A Review of Lie Detection Techniques.pdfA Review of Lie Detection Techniques.pdf
A Review of Lie Detection Techniques.pdf
 
A Review of Lie Detection Techniques
A Review of Lie Detection TechniquesA Review of Lie Detection Techniques
A Review of Lie Detection Techniques
 

More from IJECEIAES

Remote field-programmable gate array laboratory for signal acquisition and de...
Remote field-programmable gate array laboratory for signal acquisition and de...Remote field-programmable gate array laboratory for signal acquisition and de...
Remote field-programmable gate array laboratory for signal acquisition and de...
IJECEIAES
 
An efficient security framework for intrusion detection and prevention in int...
An efficient security framework for intrusion detection and prevention in int...An efficient security framework for intrusion detection and prevention in int...
An efficient security framework for intrusion detection and prevention in int...
IJECEIAES
 
A review on internet of things-based stingless bee's honey production with im...
A review on internet of things-based stingless bee's honey production with im...A review on internet of things-based stingless bee's honey production with im...
A review on internet of things-based stingless bee's honey production with im...
IJECEIAES
 
Seizure stage detection of epileptic seizure using convolutional neural networks
Seizure stage detection of epileptic seizure using convolutional neural networksSeizure stage detection of epileptic seizure using convolutional neural networks
Seizure stage detection of epileptic seizure using convolutional neural networks
IJECEIAES
 
Hyperspectral object classification using hybrid spectral-spatial fusion and ...
Hyperspectral object classification using hybrid spectral-spatial fusion and ...Hyperspectral object classification using hybrid spectral-spatial fusion and ...
Hyperspectral object classification using hybrid spectral-spatial fusion and ...
IJECEIAES
 
SADCNN-ORBM: a hybrid deep learning model based citrus disease detection and ...
SADCNN-ORBM: a hybrid deep learning model based citrus disease detection and ...SADCNN-ORBM: a hybrid deep learning model based citrus disease detection and ...
SADCNN-ORBM: a hybrid deep learning model based citrus disease detection and ...
IJECEIAES
 
Performance enhancement of machine learning algorithm for breast cancer diagn...
Performance enhancement of machine learning algorithm for breast cancer diagn...Performance enhancement of machine learning algorithm for breast cancer diagn...
Performance enhancement of machine learning algorithm for breast cancer diagn...
IJECEIAES
 
A deep learning framework for accurate diagnosis of colorectal cancer using h...
A deep learning framework for accurate diagnosis of colorectal cancer using h...A deep learning framework for accurate diagnosis of colorectal cancer using h...
A deep learning framework for accurate diagnosis of colorectal cancer using h...
IJECEIAES
 
Predicting churn with filter-based techniques and deep learning
Predicting churn with filter-based techniques and deep learningPredicting churn with filter-based techniques and deep learning
Predicting churn with filter-based techniques and deep learning
IJECEIAES
 
Automatic customer review summarization using deep learningbased hybrid senti...
Automatic customer review summarization using deep learningbased hybrid senti...Automatic customer review summarization using deep learningbased hybrid senti...
Automatic customer review summarization using deep learningbased hybrid senti...
IJECEIAES
 

More from IJECEIAES (20)

Remote field-programmable gate array laboratory for signal acquisition and de...
Remote field-programmable gate array laboratory for signal acquisition and de...Remote field-programmable gate array laboratory for signal acquisition and de...
Remote field-programmable gate array laboratory for signal acquisition and de...
 
Detecting and resolving feature envy through automated machine learning and m...
Detecting and resolving feature envy through automated machine learning and m...Detecting and resolving feature envy through automated machine learning and m...
Detecting and resolving feature envy through automated machine learning and m...
 
Smart monitoring technique for solar cell systems using internet of things ba...
Smart monitoring technique for solar cell systems using internet of things ba...Smart monitoring technique for solar cell systems using internet of things ba...
Smart monitoring technique for solar cell systems using internet of things ba...
 
An efficient security framework for intrusion detection and prevention in int...
An efficient security framework for intrusion detection and prevention in int...An efficient security framework for intrusion detection and prevention in int...
An efficient security framework for intrusion detection and prevention in int...
 
Developing a smart system for infant incubators using the internet of things ...
Developing a smart system for infant incubators using the internet of things ...Developing a smart system for infant incubators using the internet of things ...
Developing a smart system for infant incubators using the internet of things ...
 
A review on internet of things-based stingless bee's honey production with im...
A review on internet of things-based stingless bee's honey production with im...A review on internet of things-based stingless bee's honey production with im...
A review on internet of things-based stingless bee's honey production with im...
 
A trust based secure access control using authentication mechanism for intero...
A trust based secure access control using authentication mechanism for intero...A trust based secure access control using authentication mechanism for intero...
A trust based secure access control using authentication mechanism for intero...
 
Fuzzy linear programming with the intuitionistic polygonal fuzzy numbers
Fuzzy linear programming with the intuitionistic polygonal fuzzy numbersFuzzy linear programming with the intuitionistic polygonal fuzzy numbers
Fuzzy linear programming with the intuitionistic polygonal fuzzy numbers
 
The performance of artificial intelligence in prostate magnetic resonance im...
The performance of artificial intelligence in prostate  magnetic resonance im...The performance of artificial intelligence in prostate  magnetic resonance im...
The performance of artificial intelligence in prostate magnetic resonance im...
 
Seizure stage detection of epileptic seizure using convolutional neural networks
Seizure stage detection of epileptic seizure using convolutional neural networksSeizure stage detection of epileptic seizure using convolutional neural networks
Seizure stage detection of epileptic seizure using convolutional neural networks
 
Analysis of driving style using self-organizing maps to analyze driver behavior
Analysis of driving style using self-organizing maps to analyze driver behaviorAnalysis of driving style using self-organizing maps to analyze driver behavior
Analysis of driving style using self-organizing maps to analyze driver behavior
 
Hyperspectral object classification using hybrid spectral-spatial fusion and ...
Hyperspectral object classification using hybrid spectral-spatial fusion and ...Hyperspectral object classification using hybrid spectral-spatial fusion and ...
Hyperspectral object classification using hybrid spectral-spatial fusion and ...
 
Fuzzy logic method-based stress detector with blood pressure and body tempera...
Fuzzy logic method-based stress detector with blood pressure and body tempera...Fuzzy logic method-based stress detector with blood pressure and body tempera...
Fuzzy logic method-based stress detector with blood pressure and body tempera...
 
SADCNN-ORBM: a hybrid deep learning model based citrus disease detection and ...
SADCNN-ORBM: a hybrid deep learning model based citrus disease detection and ...SADCNN-ORBM: a hybrid deep learning model based citrus disease detection and ...
SADCNN-ORBM: a hybrid deep learning model based citrus disease detection and ...
 
Performance enhancement of machine learning algorithm for breast cancer diagn...
Performance enhancement of machine learning algorithm for breast cancer diagn...Performance enhancement of machine learning algorithm for breast cancer diagn...
Performance enhancement of machine learning algorithm for breast cancer diagn...
 
A deep learning framework for accurate diagnosis of colorectal cancer using h...
A deep learning framework for accurate diagnosis of colorectal cancer using h...A deep learning framework for accurate diagnosis of colorectal cancer using h...
A deep learning framework for accurate diagnosis of colorectal cancer using h...
 
Swarm flip-crossover algorithm: a new swarm-based metaheuristic enriched with...
Swarm flip-crossover algorithm: a new swarm-based metaheuristic enriched with...Swarm flip-crossover algorithm: a new swarm-based metaheuristic enriched with...
Swarm flip-crossover algorithm: a new swarm-based metaheuristic enriched with...
 
Predicting churn with filter-based techniques and deep learning
Predicting churn with filter-based techniques and deep learningPredicting churn with filter-based techniques and deep learning
Predicting churn with filter-based techniques and deep learning
 
Taxi-out time prediction at Mohammed V Casablanca Airport
Taxi-out time prediction at Mohammed V Casablanca AirportTaxi-out time prediction at Mohammed V Casablanca Airport
Taxi-out time prediction at Mohammed V Casablanca Airport
 
Automatic customer review summarization using deep learningbased hybrid senti...
Automatic customer review summarization using deep learningbased hybrid senti...Automatic customer review summarization using deep learningbased hybrid senti...
Automatic customer review summarization using deep learningbased hybrid senti...
 

Recently uploaded

Call Girls in South Ex (delhi) call me [🔝9953056974🔝] escort service 24X7
Call Girls in South Ex (delhi) call me [🔝9953056974🔝] escort service 24X7Call Girls in South Ex (delhi) call me [🔝9953056974🔝] escort service 24X7
Call Girls in South Ex (delhi) call me [🔝9953056974🔝] escort service 24X7
9953056974 Low Rate Call Girls In Saket, Delhi NCR
 
Hospital management system project report.pdf
Hospital management system project report.pdfHospital management system project report.pdf
Hospital management system project report.pdf
Kamal Acharya
 
1_Introduction + EAM Vocabulary + how to navigate in EAM.pdf
1_Introduction + EAM Vocabulary + how to navigate in EAM.pdf1_Introduction + EAM Vocabulary + how to navigate in EAM.pdf
1_Introduction + EAM Vocabulary + how to navigate in EAM.pdf
AldoGarca30
 
Standard vs Custom Battery Packs - Decoding the Power Play
Standard vs Custom Battery Packs - Decoding the Power PlayStandard vs Custom Battery Packs - Decoding the Power Play
Standard vs Custom Battery Packs - Decoding the Power Play
Epec Engineered Technologies
 

Recently uploaded (20)

Call Girls in South Ex (delhi) call me [🔝9953056974🔝] escort service 24X7
Call Girls in South Ex (delhi) call me [🔝9953056974🔝] escort service 24X7Call Girls in South Ex (delhi) call me [🔝9953056974🔝] escort service 24X7
Call Girls in South Ex (delhi) call me [🔝9953056974🔝] escort service 24X7
 
AIRCANVAS[1].pdf mini project for btech students
AIRCANVAS[1].pdf mini project for btech studentsAIRCANVAS[1].pdf mini project for btech students
AIRCANVAS[1].pdf mini project for btech students
 
Hostel management system project report..pdf
Hostel management system project report..pdfHostel management system project report..pdf
Hostel management system project report..pdf
 
kiln thermal load.pptx kiln tgermal load
kiln thermal load.pptx kiln tgermal loadkiln thermal load.pptx kiln tgermal load
kiln thermal load.pptx kiln tgermal load
 
Computer Lecture 01.pptxIntroduction to Computers
Computer Lecture 01.pptxIntroduction to ComputersComputer Lecture 01.pptxIntroduction to Computers
Computer Lecture 01.pptxIntroduction to Computers
 
Thermal Engineering Unit - I & II . ppt
Thermal Engineering  Unit - I & II . pptThermal Engineering  Unit - I & II . ppt
Thermal Engineering Unit - I & II . ppt
 
Navigating Complexity: The Role of Trusted Partners and VIAS3D in Dassault Sy...
Navigating Complexity: The Role of Trusted Partners and VIAS3D in Dassault Sy...Navigating Complexity: The Role of Trusted Partners and VIAS3D in Dassault Sy...
Navigating Complexity: The Role of Trusted Partners and VIAS3D in Dassault Sy...
 
A CASE STUDY ON CERAMIC INDUSTRY OF BANGLADESH.pptx
A CASE STUDY ON CERAMIC INDUSTRY OF BANGLADESH.pptxA CASE STUDY ON CERAMIC INDUSTRY OF BANGLADESH.pptx
A CASE STUDY ON CERAMIC INDUSTRY OF BANGLADESH.pptx
 
Computer Networks Basics of Network Devices
Computer Networks  Basics of Network DevicesComputer Networks  Basics of Network Devices
Computer Networks Basics of Network Devices
 
Hospital management system project report.pdf
Hospital management system project report.pdfHospital management system project report.pdf
Hospital management system project report.pdf
 
FEA Based Level 3 Assessment of Deformed Tanks with Fluid Induced Loads
FEA Based Level 3 Assessment of Deformed Tanks with Fluid Induced LoadsFEA Based Level 3 Assessment of Deformed Tanks with Fluid Induced Loads
FEA Based Level 3 Assessment of Deformed Tanks with Fluid Induced Loads
 
1_Introduction + EAM Vocabulary + how to navigate in EAM.pdf
1_Introduction + EAM Vocabulary + how to navigate in EAM.pdf1_Introduction + EAM Vocabulary + how to navigate in EAM.pdf
1_Introduction + EAM Vocabulary + how to navigate in EAM.pdf
 
Tamil Call Girls Bhayandar WhatsApp +91-9930687706, Best Service
Tamil Call Girls Bhayandar WhatsApp +91-9930687706, Best ServiceTamil Call Girls Bhayandar WhatsApp +91-9930687706, Best Service
Tamil Call Girls Bhayandar WhatsApp +91-9930687706, Best Service
 
Generative AI or GenAI technology based PPT
Generative AI or GenAI technology based PPTGenerative AI or GenAI technology based PPT
Generative AI or GenAI technology based PPT
 
HOA1&2 - Module 3 - PREHISTORCI ARCHITECTURE OF KERALA.pptx
HOA1&2 - Module 3 - PREHISTORCI ARCHITECTURE OF KERALA.pptxHOA1&2 - Module 3 - PREHISTORCI ARCHITECTURE OF KERALA.pptx
HOA1&2 - Module 3 - PREHISTORCI ARCHITECTURE OF KERALA.pptx
 
Bhubaneswar🌹Call Girls Bhubaneswar ❤Komal 9777949614 💟 Full Trusted CALL GIRL...
Bhubaneswar🌹Call Girls Bhubaneswar ❤Komal 9777949614 💟 Full Trusted CALL GIRL...Bhubaneswar🌹Call Girls Bhubaneswar ❤Komal 9777949614 💟 Full Trusted CALL GIRL...
Bhubaneswar🌹Call Girls Bhubaneswar ❤Komal 9777949614 💟 Full Trusted CALL GIRL...
 
Standard vs Custom Battery Packs - Decoding the Power Play
Standard vs Custom Battery Packs - Decoding the Power PlayStandard vs Custom Battery Packs - Decoding the Power Play
Standard vs Custom Battery Packs - Decoding the Power Play
 
Engineering Drawing focus on projection of planes
Engineering Drawing focus on projection of planesEngineering Drawing focus on projection of planes
Engineering Drawing focus on projection of planes
 
Orlando’s Arnold Palmer Hospital Layout Strategy-1.pptx
Orlando’s Arnold Palmer Hospital Layout Strategy-1.pptxOrlando’s Arnold Palmer Hospital Layout Strategy-1.pptx
Orlando’s Arnold Palmer Hospital Layout Strategy-1.pptx
 
data_management_and _data_science_cheat_sheet.pdf
data_management_and _data_science_cheat_sheet.pdfdata_management_and _data_science_cheat_sheet.pdf
data_management_and _data_science_cheat_sheet.pdf
 

Recognition of new gestures using myo armband for myoelectric prosthetic applications

  • 1. International Journal of Electrical and Computer Engineering (IJECE) Vol. 10, No. 6, December 2020, pp. 5694∼5702 ISSN: 2088-8708, DOI: 10.11591/ijece.v10i6.pp5694-5702 Ì 5694 Recognition of new gestures using myo armband for myoelectric prosthetic applications Jabbar Salman Hussain, Ahmed Al-Khazzar, Mithaq Nama Raheema College of Engineering, University of Kerbala, Iraq Article Info Article history: Received Feb 18, 2020 Revised May 5, 2020 Accepted May 19, 2020 Keywords: Electromyography Gestures MLPNN Multilayer perceptron neural network Myo armband Myoelectric Prosthetic Recognition ABSTRACT Myoelectric prostheses are a viable solution for people with amputations. The chal- lenge in implementing a usable myoelectric prosthesis lies in accurately recognizing different hand gestures. The current myoelectric devices usually implement very few hand gestures. In order to approximate a real hand functionality, a myoelectric prosthe- sis should implement a large number of hand and finger gestures. However, increasing number of gestures can lead to a decrease in recognition accuracy. In this work a Myo armband device is used to recognize fourteen gestures (five build in gestures of Myo armband in addition to nine new gestures). The data in this research is collected from three body-able subjects for a period of 7 seconds per gesture. The proposed method uses a pattern recognition technique based on Multi-Layer Perceptron Neural Network (MLPNN). The results show an average accuracy of 90.5% in recognizing the proposed fourteen gestures. Copyright c 2020 Insitute of Advanced Engineeering and Science. All rights reserved. Corresponding Author: Ahmed Al-Khazzar, University of Kerbala, Kerbala, Iraq. Email: ahmed.m.ahmed@uokerbala.edu.iq 1. INTRODUCTION Upper-limb myoelectric prostheses are the next generation devices aimed at helping the amputees who suffer from the loss of hand. Unlike the traditional hand prostheses which offer cosmetic or body powered lim- ited functionality to the amputee, the myoelectric prostheses use electromyography (EMG) sensors attached to the body muscles to control the prosthesis through a controller and motors that move the synthetic hand. The mechanical aspect of the prosthetics has developed vastly in recent years and many prosthetic hands with multiple movements and Degree of Freedom (DoF) are made available commercially. Examples of these pros- thetic hands are Bebionic hand [1], iLimb hand [2], Vanderbilt hand [3], the UNB hand [4], the Yale hand [5], SmartHand [6], the DLR/HIT hand [7], and the Keio hand [8]. Most of the currently developed myoelectric prostheses suffer from two problems. They either offer a limited number of hand gestures or poor recognition rate for a larger number of gestures. Since the role of a prostheses is to approximate the functionality of a real hand, there is a need for a larger number of gestures to be recognized in a myoelectric prosthesis with high recognition rate. The current paper tries to address the above problem. Developing a controller for multiple hand movements that achieves acceptable recognition accuracy is a challenge. The recognition rate and accuracy of EMG systems has been researched widely in the past [9]. According to [10], the recognition accuracy of such a system will depend mainly on a number of parameters: EMG electrode location, the signal processing methods, feature extraction techniques, and the classification algorithms. Journal homepage: http://ijece.iaescore.com
  • 2. Int J Elec & Comp Eng ISSN: 2088-8708 Ì 5695 According to [11], in surface EMG, the electrode position is vital for the detection of the signal and can affect the quality of the signal. EMG data can be different when are obtained from different nearby locations of the muscle. The use of multi-channel techniques can improve the quality of the signal with respect to the position of the electrodes. Recently, in some myoelectric prostheses, the Myo armband device has been used to collect EMG data. The Myo armband consists of eight electrodes which can mitigate the position issue. Examining the previous work shows that most of the research in the past used between two to eight EMG sensors in the forearm [12]. In regard to other parameters affecting the recognition accuracy, various feature extraction and clas- sification algorithms were researched in the previous literature, as summarized in [13]. Recognizing multiple gestures is a challenge as the recognition accuracy decreases with the increase in the number of gestures. Table 1 summarizes examples of previous literature attempting to recognize several gestures [10, 14-25]. The proposed method in this work is to use multi-layer perceptron neural network (MLPNN) com- bined with pre-processing techniques and feature extraction algorithms. In this research, an eight-channel Myo armband EMG device is used to recognize fourteen hand gestures. These are nine gestures more than the five built-in gestures the device can recognize. The novelty in this work lies in the recognition of fourteen gestures with an advanced neural network with high recognition accuracy and improving upon the results reported in the literature shown in Table 1. To the author’s knowledge, no myoelectric prosthesis is commercially available today that provides multiple gesture recognition capability. This is due to various factors such as: signifi- cant processing time of pattern recognition systems, the lack of reliability of these systems (due to electrode positioning and fatigue), and the lack of intuitive controls [26-28]. Table 1. Recognition accuracy of previous literature with multiple gestures Reference Features No. of EMG Channels Classifier No. of Subjects Time Length (seconds) No. of Gestures Recognition Accuracy (%) 10 I-EMG, MAV, MSR, 6 PCA algorithm and 5 5 6 99.6 VAR, WL, SSI, SVM classifier 11 95.6 SSC, ZC, and WA 17 95.1 14 PSD 6 ANN 12 5 9 72.9 17 63.8 15 IRMS 4 ICA,IRAM,ANN 4 10 4 90.33 16 TD, ACCC, 4 KNN, LDA, SVM 6 0.064 9 91 and SPM 17 PSD-Av 5 PNM 4 1 10 80 18 HD-sEMG 8 MK-MMD 23 3-10 22 84.6 19 Third-order AR 2 Linear Bayesian 4 1 5 90-93.5 model Coefficients, Classifier 11 83.1-95.4 MAV and MAVR 16 78.8-90.3 20 CCA 4 KNN,LDA, and 8 5-10 8 82 LIBSNM 21 MAV 64 HD 5 2 9 78.21 22 MAV,ZC, 2 KNN 30 5 4 94 SSC, and WL 23 MV, WL 4 MLE 8 1-2 8 85.7 24 MAV, AR, 4 SVM,LDA, 18 1 8 89.3 and MNF and HMM 25 MAV,ARSSC 6 SVM 5 4 5 96 ZC,WL, and RMS 13 RMS, STD, 8 MLPNN 3 5 5 99 MAX, and MIN This work RMS 8 MLPNN 3 7 14 91-94 The advance of pattern recognition techniques can potentially help recognizing multiple DoF and mapping them to myoelectric prostheses. However, recognizing gestures which include individual finger move- ments using the EMG can be a challenging task. The reason is that the EMG signal amplitude variances are small for finger movements compared to arm and wrist movements [14]. In addition, the muscles responsible for controlling the movement of fingers are located in intermediate and deep layers of forearm [27]. In order to recognize various finger movements, multiple EMG sensors are required to provide adequate data. This paper is organized as follows. Section 2 presents the structure of the Myo armband device, and the Multi-Layer Perceptron Neural Network (MLPNN). Section 3 presents data collection and the training of the MLPNN. Results are discussed in Section 4. Finally the conclusion is presented. Recognition of new gestures using myo armband for... (Jabbar Salman Hussain)
  • 3. 5696 Ì ISSN: 2088-8708 2. THEORY 2.1. Myo armband device Myo armband is a wearable device that uses eight electrode sensors to measure the EMG signals of forearm muscles. The device sends these data via Bluetooth to a computer. A special driver program is used to analyze the signals and recognize hand gestures of the user. The recognized gestures can subsequently be used by various applications. The Myo armband is also equipped with several other sensors, such as accelerometer, gyroscope, and magnetometer which collect spatial data about the gesture and can facilitate gesture recognition (these additional sensors were not utilized in this work). The device should be in direct contact with the user skin, in order to operate and can work with several operating systems such as Windows, Mac, iOS, and Android. It is powered by an ARM Cortex M4 processor and uses a rechargeable battery to operate [29-30]. 2.2. Multi-layer perceptron neural network Multi-layer perceptron neural network (MLPNN) is a feed-forward network which contains one or more invisible layers called hidden layers. The learning process is performed using a supervised method where the desired output must be known in advance to update the weights of the internal connections between the layers. This update algorithm is called the backpropagation. In this algorithm the error between the desired output and the actual output is calculated every time the input/output training data is presented. Connection weights are continuously adjusted according to the calculated error until the error gradient reaches an appro- priate small value, which indicates that the actual output is close to the target [31-34]. 3. WORK 3.1. Data collection In this work, data sets of fourteen gestures were collected from three body-able subjects using Myo armband device, as shown in Figure 1. The first five gestures (1 to 5) were selected based on the standard Myo Armband gestures. The additional nine gestures were selected based on the natural movement of the hand and fingers and some local cultural gestures. Each of the fourteen gestures were repeated twice, and the EMG signal was recorded for a period of seven seconds per gesture with a frequency of 200Hz. Table 2 represents the hand gesture categories and their description. Table 2. Hand gestures categories and the description Modes Gestures Description Relax Rest The relax state Wrist Movements Wave right Wave hand to the right Wave Left Wave hand to the left Individual Finger Movements One Finger Raise the index finger up Thumb Raise the thumb up Multi-Fingers Movement Spread Fingers Spread all fingers Fist Fist state Two Fingers Raise the index and middle fingers up Three Fingers Raise the index, middle and ring fingers up Four Fingers Raise all fingers except the thumb up Grab Grab gesture (a glass or bottle) Pinch Pinch the fingers Shake no Thumb Shake state while holding thumb inside Shake Hand Shake state with thumb up Each gesture’s data was stored as a sequence with a length varying between 1398 to 1408 EMG values (elements). A remove-silence process was applied to the data followed by removing the extra elements to create equal data sets. Furthermore, the elements in the data sets were overlapped to maximize number of samples. This yielded a data set with a dimension of 50x117 per gesture per EMG channel. Four features, namely Root Mean Square (RMS), Standard Deviation (STD), minimum (Min) and maximum (Max) values, were extracted from each gesture sequence. This process has obtained a (4x117) matrix of features per each EMG channel. For 8 channels, the result was a (32x117) features’ matrix per gesture. 80% of the obtained data were used for training (input vector), 10% were used for validation and the remaining 10% were used for testing the neural network. Int J Elec & Comp Eng, Vol. 10, No. 6, December 2020 : 5694 – 5702
  • 4. Int J Elec & Comp Eng ISSN: 2088-8708 Ì 5697 Figure 1. Representation of the fourteen gestures used in the work 3.2. Training of MLPNN For the pattern recognition, a classifier was designed using the MLPNN and based on time domain features of RMS, STD, Min and Max. The structure of the MLPNN was configured with one hidden layer to reduce processing time in the training process. The number of hidden neurons was selected based on trial and error and from experience to achieve the lowest RMS error possible. In addition, the aim was to minimize the number of training epochs in the network. Figure 2 shows the neural structure that was used for training. The classifier consisted of 32 neurons in the input layer, 30 hidden processors with tangent function, and 14 output neurons with sigmoidal function to classify 14 determined gestures. Gradient descent back-propagation was used as a training algorithm with an adaptive learning rate to minimize the risks of local minima errors. The training must be terminated if any of the stopping criteria is obtained to prevent the over fitting problem. These criteria were set to 1000 training iterations and an RMS error of 0.001. Figure 3 shows the algorithms used for training and testing of the MLPNN classifier. Recognition of new gestures using myo armband for... (Jabbar Salman Hussain)
  • 5. 5698 Ì ISSN: 2088-8708 Figure 2. Structure of neural classifier Begin Prepare the Myo Armband sensor Read the EMG data (training samples) Training MLPNN No Yes Stopping Criteria  is met? Save the trained MLPNN End Begin Read the new EMG data (test samples) Perform gesture recognition using the trained MLPNN classifier No Yes Test another  gesture? End Pre-Processing Segmentation Overlapping Feature Extraction (a) (b) Pre-Processing Segmentation Overlapping Feature Extraction Output the recognition decision Figure 3. Algorithms used for (a) the training, and (b) the testing of MLPNN classifier 4. RESULTS Figure 4 shows the performance of the network at training, validation, and test processes. It can be observed that the network was successfully trained with a validation performance of 0.001541. Figure 5 shows the error gradient for all training epochs with the RMS value of 0.00035374 at epoch 101. Number of validation checks was set to six. This validation check was achieved at epoch 101 as shown in the figure. Int J Elec & Comp Eng, Vol. 10, No. 6, December 2020 : 5694 – 5702
  • 6. Int J Elec & Comp Eng ISSN: 2088-8708 Ì 5699 Figure 4. Performance of the neural network at training, validation, and test processes Figure 5. Results of training MLPNN: the x-axis shows the number of epochs, the y-axis is the gradient of error at the top and the validation check value at bottom subfigure Figure 6 shows the real performance of the MLPNN at the end of the training and testing processes. The gesture number and the class number are represented at X-axis and Y-axis respectively. The training results provide a perfect classification with high accuracy for all of the gestures. The class number of the gesture at any testing process is defined by the largest value from all neurons at the output layer of MLPNN. For binary patterns, the output of the categorized gesture must have a value of one, and the output of the other gestures must have a value of zero. Table 3 shows the numerical representation of gesture classes. Recognition of new gestures using myo armband for... (Jabbar Salman Hussain)
  • 7. 5700 Ì ISSN: 2088-8708 Figure 6. Real performance of MLPNN at (a) training process, (b) testing process The segmentation and overlapping methods that were used in this work as part of the data pre- processing improved upon the training efficiency by increasing the number of training samples. The accuracy of the trained MLPNN is evaluated by calculating the number of correctly recognized classes in the test pro- cess. Thus the recognition rate is calculated by: RecognitionRate = (CorrectlyClassifiedSamples/AllTestSamples) ∗ 100 (1) Table 4 shows the recognition rates results for the three body able subjects. The recognition accuracy was up to 94%. Comparing these results to the previous EMG based recognition systems utilizing neural networks shows that a better recognition accuracy was achieved in this work with respect to number of gestures that can be recognized as shown in Table 5. To the authors knowledge, there were no work available using the Myo armband for recognition of multiple gestures for comparison purposes. The previous work of the same authors reported 99% of accuracy using similar neural network and 5 gestures. Hence adding 9 new gestures decreased the accuracy of the system to 90.5%. The decreased accuracy can be attributed to an increase in the number of gestures that are differentiated based on finger movements. The EMG signal amplitude variances are small for finger movements compared to arm and wrist movements which results in similar extracted features and less accurate recognition. Table 3. Assigned class numbers to the gestures Gestures Rest Spread fingers Fist Wave Right Wave Left One Finger Two Finger Three Finger Four Finger Thumb Grab Pinch Shake no Thumb Shake Hand Real O/P 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Table 4. Recognition rates of the three subjects Subject No. 1st Set Recognition Rate % 2nd Set Recognition Rate % Recognition Average % 1 89 89 89 2 89 91 90 3 91 94 92.5 Table 5. Comparison of results with other ANN classifiers in the literature Reference No. of EMG Channels No. of Subjects Time Length (seconds) No. of Gestures Recognition Accuracy (%) 14 6 12 5 9 72.9 17 63.8 15 I4 4 10 4 90.33 13 8 3 5 5 99 This work 8 3 7 14 90.5 Int J Elec & Comp Eng, Vol. 10, No. 6, December 2020 : 5694 – 5702
  • 8. Int J Elec & Comp Eng ISSN: 2088-8708 Ì 5701 5. CONCLUSION This research proposed using an MLPNN to classify and recognize additional gestures from the Myo armband device. The proposed system works by measuring the raw EMG signals from the forearm and esti- mating the muscles activity by extracting the effective time domain features such as RMS, STD, Min and Max. Then the system trains the MLPNN to classify different finger and wrist gestures. The main contribution of this work is the accurate recognition of fourteen gestures using the Myo armband device which can only recognize five gestures. The proposed pre-processing technique based on time domain feature extraction improved the training process and was effective for the recognition of gestures. The results showed an average recognition accuracy of 90.5% for all the test data. Based on these results, the proposed method can be used as a reliable classifier for different hand gestures in applications where recognizing as many gestures as possible is important such as multi-gesture prosthetic hands. The implemented system can also contribute to the advancing of the intelligent interaction between the human and instruments such as myoelectric prosthetic hands, machine control and Human Computer Interface (HCI). Future work can design and implement a manageable system for recognition of hand gestures using hardware implementation techniques such as Field Programmable Gate Array (FPGA). The error in recognition could be due to convergent gestures. The trained MLPNN may mis-classify these gestures as a result of similar features calculated from comparable raw EMG signals due to the contiguity of their associative muscles. The mis-recognized gestures vary depending on the subject under test, since each subject could perform the same gesture slightly different each time. An extension to this work can research methods to deal with these incorrect recognitions in order to improve the accuracy of the system. Finally, An MPLNN based recognition algorithm trained by back-propagation could suffer from local minima or high computational cost. Future work needs to examine other intelligent algorithms such as Adaptive Neuro Fuzzy Inference System (ANFIS) for gesture recognition. ACKNOWLEDGEMENT This work is carried out by Intelligent Prosthetics Research Group (IPRG) at College of Engineering, University of Kerbala in 2019. REFERENCES [1] Bebionic Hand, Ottobock, [Online] Available: https://www.ottobockus.com/prosthetics/upper-limb- prosthetics/solution-overview/bebionic-hand. [2] I-Limb Ultra, Ossur, [Online] Available: https://www.ossur.com/en-us/prosthetics/arms/i-limb-ultra. [3] S. A. Dalley, et al. “Design of a multifunctional anthropomorphic prosthetic hand with extrinsic actua- tion,” IEEE/ASME Transactions on Mechatronics, vol. 14, no. 6, pp. 699-706, 2009. [4] Y. Losier, et al., “An overview of the UNB hand system,” Proc. of the Myoelectric Controls/Powered Prosthetics Symp., 2011. [5] J. T. Belter, and A. M. Dollar, “Novel differential mechanism enabling two DoF from a single actuator: Application to a prosthetic hand,” IEEE 13th Int. Conf. on Rehabilitation Robotics, pp. 1-6, 2013. [6] C. Cipriani, M. Controzzi, and M.C Carrozza, “The SmartHand transradial prosthesis,” Journal of neuro- engineering and rehabilitation, vol. 8, no. 1, 2011. [7] H. Liu, et al., “Multisensory five-finger dexterous hand: The DLR/HIT Hand II,” IEEE/RSJ international conference on intelligent robots and systems, pp. 3692-3697, 2008. [8] Y. Kamikawa, and T. Maeno, “Underactuated five-finger prosthetic hand inspired by grasping force dis- tribution of humans,” IEEE/RSJ Int. Conf. on Intelligent Robots and Systems, pp. 717-722, 2008. [9] N. E. Krausz, and R. A. Rorrer, “Design and fabrication of a six degree-of-freedom open source hand,” IEEE Transactions on Neural Systems and Rehabilitation Engineering, vol. 24, no. 5, pp. 562-572, 2015. [10] Z. Chen, et al., “Hand gestures recognition from multi-channel forearm EMG signals,” International Conference on Cognitive Systems and Signal Processing, pp. 119-125, 2016. [11] L. Mesin, R. Merletti, and A. Rainoldi, “Surface EMG: the issue of electrode location,” Journal of Elec- tromyography and Kinesiology, vol. 19, no. 5, 2009. [12] X. Zhang, et al., “A framework for hand gesture recognition based on accelerometer and EMG sensors,” IEEE Transactions on Systems, Man, and Cybernetics-Part A: Systems and Humans, vol. 41, no. 6, pp. 1064-1076, 2011. Recognition of new gestures using myo armband for... (Jabbar Salman Hussain)
  • 9. 5702 Ì ISSN: 2088-8708 [13] A.M. AlKhazzar and M.N. Raheema, “EMG Signal Classification Using Radial Basis Function Neural Network,” Third Scientific Conference of Electrical Engineering (SCEE), pp. 180-185, 2018. [14] T. Jiralerspong, et al., “Experimental study of real-time classification of 17 voluntary movements for multi-degree myoelectric prosthetic hand,” Applied Sciences, vol. 7, no. 1, 2017. [15] G. R. Naik, D. K. Kumar, and S. P. Arjunan, “Multi modal gesture identification for HCI using surface EMG,” Proc. of the 12th Int. conf. on entertainment and media in the ubiquitous era, pp. 90-94, 2008. [16] X. Chen, and Z.J. Wang, “Pattern recognition of number gestures based on a wireless surface EMG system,” Biomedical Signal Processing and Control, vol. 8, no. 2, pp. 184-192, 2013. [17] M.C. Castro, S.P. Arjunan, and D.K. Kumar, “Selection of suitable hand gestures for reliable myoelectric human computer interface,” Biomedical engineering online, vol. 14, no. 1, 2015 [18] Y. Du, et al., “Surface EMG-based inter-session gesture recognition enhanced by deep domain adapta- tion,” Sensors, vol. 17, no. 3, 2017. [19] X. Chen, et al., “Hand gesture recognition research based on surface EMG sensors and 2D- accelerometers,” 11th IEEE International Symposium on Wearable Computers, pp. 11-14, 2007. [20] N. K. Rami, “Correlation Analysis of Electromyogram Signals for Multiuser Myoelectric Interfaces,” IEEE tran. on neural systems and rehabilitation eng., vol. 22, no. 4, pp. 745-755, 2014. [21] A. Moin, et al., “Analysis of Contraction Effort Level in EMG-Based Gesture Recognition Using Hyper- dimensional Computing,” IEEE biomedical circuits and systems conf., 2019. [22] W.T. Shi, et al., “A bionic hand controlled by hand gesture recognition based on surface EMG signals: A preliminary study,” Biocybernetics and Biomedical Engineering, vol. 38, no. 1, pp. 126-135, 2018. [23] K. Rhee, and H.C. Shin, “Electromyogram-based hand gesture recognition robust to various arm pos- tures,” International Journal of Distributed Sensor Networks, vol. 14, no. 7, 2018. [24] W. Geng, Du, Y., Jin, W., Wei, W., Hu, Y., and Li, J, “Gesture recognition by instantaneous surface EMG images,” Scientific reports, vol. 6, 2016. [25] X. Luo, et al., “Synergistic Myoelectrical Activities of Forearm Muscles Improving Robust Recognition of Multi-Fingered Gestures,” Sensors, vol. 19, no. 3, pp. 610, 2019. [26] A.D. Roche, et al., “Prosthetic myoelectric control strategies: a clinical perspective,” Current Surgery Reports, vol. 2, no. 3, pp. 44, 2014. [27] R. L. Drake, A. W. Vogl, and A. W. Mitchell, “Gray’s anatomy for students 2nd ed.,” Churchill Living- stone, 2009. [28] A.H. Al-Timemy, et al., “Improving the performance against force variation of EMG controlled multi- functional upper-limb prostheses for transradial amputees,” IEEE Transactions on Neural Systems and Rehabilitation Engineering, vol. 24, no. 6, pp. 650-661, 2015. [29] S. Rawat, S. Vats and P. Kumar, ”Evaluating and exploring the MYO ARMBAND,” Int. Conf. System Modeling and Advancement in Research Trends (SMART), pp. 115-120, 2016. [30] “Myo Gesture Control Armband tech specs,” [Online] Available: https://support.getmyo.com/ hc/en- us/articles/202648103-Myo-Gesture-Control-Armband-tech-specs. [31] N. F. Ebecken, “An overview on the use of neural networks for data mining tasks,” Journal of the Brazilian Neural Network Society, vol. 9, no. 3, pp. 202-212, 2011. [32] G. Panchal, et al., “Behaviour analysis of multilayer perceptrons with multiple hidden neurons and hidden layers,” Int. Journal of Computer Theory and Eng., vol. 3, no. 2, pp. 332-337, 2011. [33] M.A. Hossain, et al., “Implementation Of Back-Propagation Neural Network For Isolated Bangla Speech Recognition,” Int. Journal of Information Sciences and Techniques (IJIST), vol. 3, no. 4, 2013. [34] M.C. Popescu, et al., “Multilayer perceptron and neural networks,” Wseas Transactions on Circuits and Systems, vol. 8, no. 7, pp. 579-588, 2009. Int J Elec & Comp Eng, Vol. 10, No. 6, December 2020 : 5694 – 5702