2024 4th International Conference of Science and Information Technology in Smart Administration (ICSINTESA)

LabVIEW Integrated with Python Machine Learning

for DC Motor Drive
2024 4th International Conference of Science and Information Technology in Smart Administration (ICSINTESA) | 979-8-3503-7611-1/24/$31.00 ©2024 IEEE | DOI: 10.1109/ICSINTESA62455.2024.10748086

1st Mana Poyai 2nd Dechrit Maneetham 3rd Petrus Sutyasadi

Department of Mechatronics Department of Mechatronics Department of Mechatronics
Engineering, Faculty of Technical Engineering, Faculty of Technical Vocational Faculty Sanata Dharma
Education, Rajamangala University of Education, Rajamangala University of University
Technology Thanyaburi Technology Thanyaburi Yogyakarta, Indonesia
Pathumthani, Thailand Pathumthani, Thailand Email: [email protected]
Email: [email protected] Email: [email protected]

Abstract— This study presents a new technique that

inside data. It’s a branch of artificial intelligence (AI), allows
integrates LabVIEW and Python to enhance the control of DC systems to autonomously learn and refine their performance
motor drives through the utilization of machine learning based on experience, eliminating the need for human
methods. The objective of our research is to leverage intervention. This technology enhances the predictive
LabVIEW's graphical programming interface in combination capabilities of software applications without the necessity for
with the Scikit-learn Python package to train a Polynomial explicit programming. Fundamentally, machine learning
Regression model. The goal is to forecast Pulse Width focuses on creating algorithms that analyze input data and
Modulation (PWM) values at desired Rotations per Minute make predictions using statistical methods. In recent years,
(RPM) of a DC motor. The accuracy and reliability of the
machine learning has become highly popular, with its
trained model are evaluated using performance assessment
metrics such as the R-squared (R2) coefficient of determination
algorithms being utilized in diverse areas such as object
and the Root Mean Square Error (RMSE). The R2 score assesses detection, pattern recognition, text interpretation, and various
the precision of the model's fit, whereas the RMSE quantifies research disciplines. The main aim of machine learning is to
the model's predictive capability by comparing observed and teach computers to leverage data to address specific problems
predicted data. This research also includes a comparison efficient [5]
between the machine learning technique for DC motor drive and In recent years, Machine Learning (ML) algorithms have
conventional PID control. The findings illustrate the successful been increasingly implemented, though some solutions may
integration of LabVIEW and Python in managing DC motor simply replace one problem with another. Common types of
drives, achieving low overshooting and faster attainment of the machine learning include supervised learning, unsupervised
set RPM point compared to classical PID control. This
emphasizes machine learning's capacity to enhance control
learning, and reinforcement learning [6].
strategies; the application of Polynomial Regression yields Supervised learning involves using a dataset comprised of
promising results, demonstrating its effectiveness in predicting pre-labeled target data input variables (training data) [7]. This
PWM and RPM values with an R2 value ranging from 96% to method generates a mapping function to associate inputs with
98% and RMSE within 19. the desired outputs based on these input variables. The model
parameters are adjusted until it reaches a satisfactory level of
Keywords—Labview, Python, Scikit-learn, Polynomial Linear accuracy with the training data. This type of learning allows for
Regression, DC motor Drive, EasyEda, Machine Learning, predictions based on a known dataset (training dataset), which
RMSE, R2, PID
includes input variables (X) and output/response variables (Y).
I. INTRODUCTION These variables are essential for developing a model that
predicts the value of the response variable (Y) for new, unseen
The incorporation of various technologies has played a data. In supervised learning, an algorithm is designed to learn
crucial role in enhancing the capabilities of engineering the mapping function from the input variables (X) to the
systems in recent years. The combination of hardware and output/response variable (Y). The mapping function can be
software, along with the utilization of machine learning
expressed as:
techniques, has created opportunities to improve the
= ( )
effectiveness and capability of different applications. Our
project aims to integrate LabVIEW [1] and Python [2] to (1)
effectively drive and control DC motors using an intelligent
machine learning framework. DC motors are essential Unsupervised learning utilizes only training data without
components in a wide range of industrial applications, associated outcome data, meaning the data is not pre-classified.
including robotics and automation systems [3]. Our work This approach identifies existing patterns or clusters within
intends to leverage the benefits of LabVIEW, a versatile datasets without the need for supervision. The model
graphical programming environment, and Python, a autonomously discovers information to produce output,
sophisticated language extensively used in data science and primarily dealing with unlabeled data. One of the most
machine learning [4], to create a control system based on a common methods in unsupervised learning is cluster analysis,
machine learning approach. The control system employs which is used to identify structures or patterns within a pool of
machine learning techniques to fine-tune and verify its unclassified data [8].
performance using statistical tools such as R² for machine Reinforcement learning involves conditioning a machine to
learning optimization, and RMSE for comparing predictions take actions based on specific decisions, resulting in rewards
with actual outcomes. Additionally, the performance of the or feedback. The machine is programmed to draw from past
machine learning approach is compared with the experiences to identify the most beneficial actions. This
conventional PID approach. method enables the machine to take appropriate actions in
II. LITERATURE REVIEW given situations to maximize rewards. Learning occurs through
interaction with the environment by agents or learners, who
A. Machine Learning Techniques receive rewards for correct actions and penalties for incorrect
Machine learning has progressed as a groundbreaking ones. The agent learns without human intervention by aiming
field, offering powerful tools to depict complex relationships to maximize rewards and minimize penalties. This system of
Authorized licensed use limited to: Rajamangala Univ of Technology Thanyaburi provided by UniNet. Downloaded on March 03,2025 at 01:45:02 UTC from IEEE Xplore. Restrictions apply.
979-8-3503-7611-1/24/$31.00 ©2024 IEEE 358
 rewards and punishments effectively trains the algorithms [9]. duty cycle is defined as the ratio of the duration in which the
signal is in the "on" state (high voltage) to the overall period of
B. Algorithm of Machine Learning the signal. Pulse Width Modulation (PWM) is a technique that
Decision tree algorithms are primarily utilized for allows for precise control of the average power output to a DC
classification problems. They divide attributes into two or motor by adjusting the duty cycle. Pulse Width Modulation
more classes based on their values. Each tree consists of nodes (PWM) enables speed control by adjusting the proportion of
and branches, where each node represents an attribute of the time the signal is on compared to the total time of the signal
group, and each branch corresponds to a specific value of that [19]. As the duty cycle increases, the motor experiences a
attribute [10]. greater average voltage, leading to an increase in speed. On the
Artificial neural networks model themselves after the other hand, a decreased duty cycle results in a decrease in the
configuration of biological neurons and operate using a average voltage and a reduction in the motor speed. This
supervised approach to machine learning. They consist of system enables seamless and uninterrupted speed regulation in
artificial neurons interconnected with weighted connections DC motor drives.
between units. These networks are recognized for their
capability in parallel distributed computing [11]. III. RESEARCH METHODOLOGY
Data Clustering Algorithms partition objects into different A. Hardware Design
groups or clusters based on similarities among items within
The web-based software EasyEDA [20] was employed to
each subset. It is an unsupervised learning method commonly create a schematic and PCB, integrating components such as
referred to as clustering or hierarchical and network a controller (ESP8266), opto-coupler isolator, IGBT drive,
partitioning techniques [12]. power regulator, and various discrete elements. This board is
Regression algorithms make predictions by modeling the designed to drive the DC motor and record RPM speed for
relationship between variables using an error measure, machine learning data collection. The schematic of this board
predicting continuously varying values. This statistical tool is illustrated in Fig. 1 (a), utilizing the ESP8266 controller
helps identify the relationship between dependent and embedded with C code firmware. It functions as the bridge
independent variables and is used to model a target value that channel interfacing through the USB port between the
relies entirely on independent predictors. Regression analysis LabVIEW-Python-based PC software, the DC motor drive,
is employed to find correlations between two variables, and the encoder reader. The board utilizes the IGBT as the
though it cannot determine cause-and-effect relationships. switching drive for the DC motor and an opto-coupler to
Unlike classification algorithms, which categorize output isolate the microcontroller drive signal from the
variables, regression focuses on predicting continuous electromagnetic induction generated by the DC motor.
outcomes [13]. Voltage regulators are employed to regulate the power supply
Polynomial regression [14] is a type of regression from the DC motor to match with the supply of the ESP8266
algorithm which has versatile technique that is particularly microcontroller, opto-coupler, and other components. The
valuable for assessing the relationship between data, functionality of this PCB includes an RPM reading feature
through the encoder, facilitating the measurement of the DC
especially when implemented using the Scikit-learn
motor speed for data collection in the machine learning
framework. This approach is highly significant in numerous
training module. The PCB shown in Fig.1 (b) is generated
disciplines [15]. from this schematic using the same EasyEDA software, with
the 3D view shown in Fig. 1 (c). The assembled PCB,
C. Polynomial Regression Fundamentals featuring components for a complete PCBA for the DC motor
Polynomial regression, an extension of linear regression, drive with data collection for machine learning, is depicted in
is adept at capturing intricate nonlinear relationships. The Fig. 1 (d).
literature extensively covers its fundamental principles,
showcasing its capability to model data with higher degrees of
complexity [16]. The general form of the polynomial
regression equation is expressed as a polynomial function of
the independent variable, allowing researchers to flexibly
adapt models to the shape of the data.
(a) (b)
D. Scikit-learn a Robust Toolkit for Polynomial Regression
Scikit-learn, a widely used machine learning library in
Python, provides a comprehensive set of tools for
implementing polynomial regression models. Researchers
have embraced Scikit-learn for its efficient implementation
and integration capabilities, enabling the seamless (c) (d)
deployment of polynomial regression in diverse applications Fig. 1 (a) Schematic of DC Motor Drive,
[17]. (b) PCB layout of DC Motor Drive , (c) PCB (3D) of DC Motor
Drive, (d) PCBA of DC Motor Drive
E. Pulse Width Modulation (PWM) for DC Motor Drives
Pulse Width Modulation (PWM) is a vital technology B. Software System Design
extensively used in the realm of DC motor drives to control LabVIEW is primarily used as the graphical user interface
the velocity and orientation of direct current (DC) motors [18]. (GUI) for this system, sending PWM signals to the ESP8266
This technology is crucial for achieving accurate control over controller to control the IGBT, thereby turning the DC motor
motor performance, enabling efficient and dynamic on and off. Additionally, LabVIEW is employed to read the
modifications in a wide range of applications, including RPM speed of the DC motor through the encoder. The
robotics and industrial automation. Pulse Width Modulation interfacing between the DC motor drive board and the PC is
(PWM) functions by swiftly alternating a signal between an achieved through the USB port. The dataset of PWM and RPM
active state and an inactive one at a consistent frequency. The values is recorded in a text file for further processing in
Authorized licensed use limited to: Rajamangala Univ of Technology Thanyaburi provided by UniNet. Downloaded on March 03,2025 at 01:45:02 UTC from IEEE Xplore. Restrictions apply.
359
 Python's text-based coding software. This Python software is C. Polynomial Regression
designed to invoke the Scikit-learn library's train module, The train model is used the polynomial method to find the
establishing the relationship between PWM (%) and RPM, relation between the input and output as the equation below.
with the R-Squared value calculation for the model's
predictive accuracy. Then, after the LabVIEW software is
triggered by the user to predict the PWM (%) value needed to = #$ + #& + #' '
+. . . +#( (
(2)
be sent to the DC motor drive board for achieving the desired
RPM, LabVIEW will call the Python to predict value. This
function predicts the PWM(%) value required for the Where is the predicted output, is the input variable and
specified RPM using the trained model previously executed. #$ , #& , … , #( are the coefficients.
In addition, this LabVIEW-Python integrated software also
includes a test mode to collect the data for evaluating the D. RMSE Calculation for Model Evaluation
trained model using the Root-mean-square deviation (RMSE) The Root Mean Square Error (RMSE) is used to objectively
value. This value represents the difference in error between evaluate the accuracy of machine learning models by comparing
the trained model and the real values generated from the their predictions to the actual values. The equation for Root
predict function. The overview flow diagram of this software Mean Square Error (RMSE) is provided as follows.
system is shown as the Fig. 2.

&
+,-. = / ∑07=&( 234526,7 − 9:;<734,7 )
' (3)
0

Where > is the number of observation, 234526,7 is the actual

value, 9:;<734,7 is predict value.

IV. RESULTS AND DISCUSSIONS

A. Machine Learning DC Motor Train and Drive

The LabVIEW-based software is executed to send the
PWM value as a percentage to the ESP8266, which drives the
IGBT in PWM mode to generate power for the DC motor. It
also records the RPM value of the DC motor sent back from
the encoder along with the PWM (%) value applied to the DC
motor. The X-Y graph illustrating the relationship between the
PWM (%) and RPM values, plotted by LabVIEW, is shown in
Fig. 4. This graph demonstrates that the PWM (%) and RPM
values exhibit unique characteristics under different
conditions. For this experiment, two models of DC motors are
used, 25GA371 (12VDC, 1000 rpm) and GA25370 (6VDC,
1360 rpm), under two conditions: no load and with load. The
systems used are shown in Fig. 5.
Fig. 2 Software Flow Diagram

The interface between LabVIEW and Python is established

through the CMD tool as shown in Fig. 3 (a) and the user
interface is shown as Fig. 3 (b), which calls the Python script
and facilitates the interchange of data with a text-based file.

(a) (b)

(a)

(c) (d)
Fig. 4 The X-PWM(%), Y-RPM graph plotted by LabVIEW
shows the relationship between the PWM(%) and the RPM
values read back from the encoder : (a) 25GA371 DC Motor
without load, (b) 25GA371 DC Motor with load, (c) GA25370 DC
Motor without load, (d) GA25370 DC Motor with load

(b)

Fig. 3 (a) Labview Call Python Script

(b) Labview User Interface
Authorized licensed use limited to: Rajamangala Univ of Technology Thanyaburi provided by UniNet. Downloaded on March 03,2025 at 01:45:02 UTC from IEEE Xplore. Restrictions apply.
360
 (a) (b)

Fig. 5 (a) 25GA371 DC Motor without load, (b) 25GA371 DC

Motor with load, (c) GA25370 DC Motor without load, (d)
GA25370 DC Motor with load

Then, after recording the PWM (%) and RPM values of

the DC motor, the Python code is used to train the model with
a polynomial degree from 1 to 4 from the Scikit-learn library.
The fitted model is plotted as shown in Fig. 6 to 9, along with
the calculation of the R2 value. Which, it seems that a (c) (d)
polynomial of degree 4 fits well, as indicated by the high R2 Fig.8 Polynomial plot degree 3 with trend line between RPM
score [21] at around 96 to 98%. And, the summary of R2 is and PWM (%), (a) 25GA371 no load, (b) 25GA371 with load, (c)
shown in the Table I. GA25-370 no load, (d) GA25-370 with load

(a) (b) (a) (b)

(c) (d) (c) (d)

Fig.6 Polynomial plot degree 1 with trend line between Fig.9 Polynomial plot degree 4 with trend line between RPM
RPM and PWM (%), (a) 25GA371 no load, (b) 25GA371 with and PWM (%), (a) 25GA371 no load, (b) 25GA371 with load, (c)
load, (c) GA25-370 no load, (d) GA25-370 with load GA25-370 no load, (d) GA25-370 with load

TABLE I. R2 SUMMARY OF POLYNOMIAL DEGREE 1 TO 4

Polynomial Degree
System Condition
1 2 3 4

25GA371 No load 87% 95% 95% 96%

(a) (b) 25GA371 No load 85% 94% 98% 98%
GA25-370 With load 95% 96% 96% 96%
GA25-370 With load 92% 92% 96% 98%

And, after completing the training, the test command from

LabVIEW is executed to compare the predicted values with
the actual values and observe the error, which is shown as
Table II to Table V and the Root-mean-square error (RMSE)
(c) (d) [22] value is calculated.
Fig.7 Polynomial plot degree 2 with trend line between
RPM and PWM (%), (a) 25GA371 no load, (b) 25GA371 with
load, (c) GA25-370 no load, (d) GA25-370 with load

Authorized licensed use limited to: Rajamangala Univ of Technology Thanyaburi provided by UniNet. Downloaded on March 03,2025 at 01:45:02 UTC from IEEE Xplore. Restrictions apply.
361
 TABLE II. RMSE CALCULATION BETWEEN DESIRE RPM VS. ACTUAL RPM B. Machine Learning DC Motor Drive Comparing with
25GA371 NO LOAD Conventional PID Control

PWM Actual RPM

Desire
(%) Average SD Diff In the present study, the DC motor drive system, which
RPM 1 2 3
Predict incorporates the machine learning concepts discussed in
400 36.85 430 435 441 435 5 35 the previous experiment from the earlier section, is
500 40.78 489 492 494 492 3 -8 meticulously compared with a traditional Proportional-
600 45.18 593 594 595 594 1 -6 Integral-Derivative (PID) control approach [24]. This
700 50.38 683 683 686 684 2 -16
comparison is conducted under identical conditions,
specifically using the 25GA371 motor under no load. The
800 56.62 780 780 781 780 1 -20
performance evaluation involves capturing the revolutions
900 64.09 893 894 896 894 1 -6 per minute (RPM) of the motor over a specified time
RMSE 18 interval, immediately after the command is issued from
the PC to achieve a particular RPM. The resulting data, as
TABLE III. RMSE CALCULATION BETWEEN DESIRE RPM VS. ACTUAL RPM illustrated in Fig. 12, reveals that the motor drive
25GA371 WITH LOAD employing the machine learning method exhibits a
significantly lower overshoot in RPM compared to the
Desire
PWM Actual RPM conventional PID control. Furthermore, the machine
(%) Average SD Diff learning-based drive achieves the desired RPM in a
RPM 1 2 3
Predict
noticeably shorter duration. This underscores the
400 29.24 390 406 400 399 8 -1
enhanced efficiency and responsiveness of the machine
500 30.54 501 513 495 503 9 3 learning approach over the conventional PID method in
600 33.21 600 611 600 604 6 4 controlling the DC motor drive.
700 39.30 710 719 705 711 7 11
800 51.28 801 798 795 798 3 -2
900 71.98 920 929 950 933 16 33
RMSE 14

TABLE IV. RMSE CALCULATION BETWEEN DESIRE RPM VS. ACTUAL RPM
GA25-370 NO LOAD

PWM Actual RPM

Desire
(%) Average SD Diff
RPM 1 2 3
Predict
400 44.38 397 392 403 397 6 -3
500 48.92 501 491 509 501 9 1 (a)
600 54.90 601 600 598 600 2 0
700 62.72 717 707 707 710 6 10
800 72.32 795 793 797 795 2 -5
900 83.17 921 897 899 906 14 6
RMSE 5

TABLE V. RMSE CALCULATION BETWEEN DESIRE RPM VS. ACTUAL RPM

GA25-370 WITH LOAD

PWM Actual RPM

Desire
(%) Average SD Diff
RPM 1 2 3
Predict
(b)
400 50.59 450 416 414 427 20 27
500 50.40 473 473 501 482 16 -18
Fig. 12 Graph of RPM speed on y-axis and time (sec) on x-
600 51.78 560 610 599 590 26 -10 axis (a) Conventional PID control for DC motor drive (b)
Machine Learning control for DC motor drive
700 55.93 674 712 679 688 21 -12
800 63.24 769 802 767 779 20 -21
900 73.26 929 915 922 922 7 22
RMSE 19

Authorized licensed use limited to: Rajamangala Univ of Technology Thanyaburi provided by UniNet. Downloaded on March 03,2025 at 01:45:02 UTC from IEEE Xplore. Restrictions apply.
362
 Science, Electric Engineering and Computer (MEC), Jilin, China, 2011,
V. CONCLUSION pp. 1143-1146, doi: 10.1109/MEC.2011.6025669.
[10] A. Navada, A. N. Ansari, S. Patil and B. A. Sonkamble, "Overview of
This study successfully demonstrated the combination use of decision tree algorithms in machine learning," 2011 IEEE
of LabVIEW and Python to enhance the capabilities of DC Control and System Graduate Research Colloquium, Shah Alam,
motor drive control. Using the Scikit-learn Python package, we Malaysia, 2011, pp. 37-42, doi: 10.1109/ICSGRC.2011.5991826.
employed polynomial regression as a machine learning [11] J. J. Hopfield, "Artificial neural networks," in IEEE Circuits and
technique [23] to predict the PWM (Pulse Width Modulation) Devices Magazine, vol. 4, no. 5, pp. 3-10, Sept. 1988, doi:
10.1109/101.8118.
percentage command required to achieve the desired speed
[12] J. Oyelade et al., "Data Clustering: Algorithms and Its
(RPM) of a DC motor. This approach was compared with Applications," 2019 19th International Conference on Computational
conventional PID control, which typically has higher overshoot Science and Its Applications (ICCSA), St. Petersburg, Russia, 2019, pp.
and requires more time to reach the target RPM value. The 71-81, doi: 10.1109/ICCSA.2019.000-1.
conventional method also involves complex fine-tuning of P, I, [13] D. Kinaneva, G. Hristov, P. Kyuchukov, G. Georgiev, P. Zahariev and
R. Daskalov, "Machine Learning Algorithms for Regression Analysis
and D values to reduce overshoot and quickly reach the RPM and Predictions of Numerical Data," 2021 3rd International Congress
set point. The machine learning concept introduced in this on Human-Computer Interaction, Optimization and Robotic
paper addresses these issues. The smooth interface between Applications (HORA), Ankara, Turkey, 2021, pp. 1-6, doi:
LabVIEW and Python enabled rapid data transmission and 10.1109/HORA52670.2021.9461298.
model training, signifying a significant advancement in control [14] H. Li and S. Yamamoto, "Polynomial regression based model-free
predictive control for nonlinear systems," 2016 55th Annual
system design. Polynomial regression yielded positive results Conference of the Society of Instrument and Control Engineers of
in predicting the PWM % value. The accuracy of the model was Japan (SICE), Tsukuba, Japan, 2016, pp. 578-582, doi:
evaluated by comparing the observed and predicted values, 10.1109/SICE.2016.7749264.
utilizing the Root Mean Square Error (RMSE). The attained [15] C. -p. Hwang, M. -S. Chen, C. -M. Shih, H. -Y. Chen and W. K. Liu,
RMSE is within 19 serves as a benchmark for assessing the "Apply Scikit-Learn in Python to Analyze Driver Behavior Based on
OBD Data," 2018 32nd International Conference on Advanced
predictive capabilities of our integrated system. This study Information Networking and Applications Workshops (WAINA),
establishes a solid foundation for the partnership between Krakow, Poland, 2018, pp. 636-639, doi:
LabVIEW and Python in the realm of machine learning-driven 10.1109/WAINA.2018.00159.
control systems. [16] C. Shang and Z. Li, "Research on the Change Rule of the Proportion of
the Added Value of Circulation Industry in GDP -- Based on nonlinear
ACKNOWLEDGMENT regression model and logarithmic model," 2020 Management Science
Informatization and Economic Innovation Development Conference
Rajamangala University of Technology Thanyaburi (MSIEID), Guangzhou, China, 2020, pp. 534-537, doi:
provided tremendous help by offering research facilities for 10.1109/MSIEID52046.2020.00108.
this study. We would like to express our profound [17] T. Bakibayev and A. Kulzhanova, "Common Movement Prediction
appreciation to all the individuals and entities who made using Polynomial Regression," 2018 IEEE 12th International
contributions to this research. Conference on Application of InformaT. Bakibayev and A.
Kulzhanova, "Common Movement Prediction using Polynomial
Regression," 2018 IEEE 12th International Conference on Application
of Information and Communication Technologies (AICT), Almaty,
REFERENCES Kazakhstan, 2018, pp. 1-4, doi: 10.1109/ICAICT.2018.8747047.
[1] R. Jamal and L. Wenzel, "The applicability of the visual programming [18] H. F. Weber, "Pulse-Width Modulation DC Motor Control," in IEEE
language LabVIEW to large real-world applications," Proceedings of Transactions on Industrial Electronics and Control Instrumentation,
Symposium on Visual Languages, Darmstadt, Germany, 1995, pp. 99- vol. IECI-12, no. 1, pp. 24-28, March 1965, doi:
106, doi: 10.1109/VL.1995.520791. 10.1109/TIECI.1965.229545.
[2] F. Dubosson, S. Bromuri and M. Schumacher, "A Python Framework [19] Z. Xu and Y. Kang, "PWM speed DC motor drive power design," 2016
for Exhaustive Machine Learning Algorithms and Features 31st Youth Academic Annual Conference of Chinese Association of
Evaluations," 2016 IEEE 30th International Conference on Advanced Automation (YAC), Wuhan, China, 2016, pp. 419-423, doi:
Information Networking and Applications (AINA), Crans-Montana, 10.1109/YAC.2016.7804930.
Switzerland, 2016, pp. 987-993, doi: 10.1109/AINA.2016.160.
[20] N. Sri Sai et al., "Efficient design of 8×8×8 LED Cube with Low
[3] H. Yu, "Application of in DC Motor Control Method in Intelligent Power consumption using Arduino UNO," 2020 11th International
Vehicles and Robots," 2022 IEEE 5th International Conference on Conference on Computing, Communication and Networking
Automation, Electronics and Electrical Engineering (AUTEEE), Technologies (ICCCNT), Kharagpur, India, 2020, pp. 1-6, doi:
Shenyang, China, 2022, pp. 1018-1021, doi: 10.1109/ICCCNT49239.2020.9225332.
10.1109/AUTEEE56487.2022.9994475.
[21] C. Reyes, T. Hilaire, S. Paul and C. F. Mecklenbräuker, "Evaluation of
[4] M. Kaur, A. Kumar Shukla and S. Kaur, "An Introduction to Machine the root mean square error performance of the PAST-Consensus
Learning in a Nutshell," 2021 10th International Conference on algorithm," 2010 International ITG Workshop on Smart Antennas
System Modeling & Advancement in Research Trends (SMART), (WSA), Bremen, Germany, 2010, pp. 156-160, doi:
MORADABAD, India, 2021, pp. 17-22, doi: 10.1109/WSA.2010.5456452.
10.1109/SMART52563.2021.9676315.
[22] N. Shabbir, R. Amadiahangar, H. A. Raja, L. Kütt and A. Rosin,
[5] S. Angra and S. Ahuja, "Machine learning and its applications: A "Residential Load Forecasting Using Recurrent Neural Networks,"
review," 2017 International Conference on Big Data Analytics and 2020 IEEE 14th International Conference on Compatibility, Power
Computational Intelligence (ICBDAC), Chirala, Andhra Pradesh, Electronics and Power Engineering (CPE-POWERENG), Setubal,
India, 2017, pp. 57-60, doi: 10.1109/ICBDACI.2017.8070809. Portugal, 2020, pp. 478-481, doi: 10.1109/CPE-
[6] P. P. Shinde and S. Shah, "A Review of Machine Learning and Deep POWERENG48600.2020.9161565.
Learning Applications," 2018 Fourth International Conference on [23] K. A. Ismail and M. A. Abd El Ghany, "High Performance Machine
Computing Communication Control and Automation (ICCUBEA), Learning Models for Functional Verification of Hardware Designs,"
Pune, India, 2018, pp. 1-6, doi: 10.1109/ICCUBEA.2018.8697857. 2021 3rd Novel Intelligent and Leading Emerging Sciences Conference
[7] Abhishek, A. Dhankar and N. Gupta, "A Systematic Review of (NILES), Giza, Egypt, 2021, pp. 15-18, doi:
10.1109/NILES53778.2021.9600502.
Techniques, Tools and Applications of Machine Learning," 2021
Third International Conference on Intelligent Communication [24] M. A. Taut, G. Chindris and D. Pitică, "PID Algorithm used for DC
Technologies and Virtual Mobile Networks (ICICV), Tirunelveli, Motor Control," 2018 IEEE 24th International Symposium for Design
India, 2021, pp. 764-768, doi: 10.1109/ICICV50876.2021.9388637. and Technology in Electronic Packaging (SIITME), Iasi, Romania,
[8] M. Usama et al., "Unsupervised Machine Learning for Networking: 2018, pp. 365-372, doi: 10.1109/SIITME.2018.8599230.
Techniques, Applications and Research Challenges," in IEEE Access,
vol. 7, pp. 65579-65615, 2019, doi: 10.1109/ACCESS.2019.2916648.
[9] W. Qiang and Z. Zhongli, "Reinforcement learning model, algorithms
and its application," 2011 International Conference on Mechatronic
Authorized licensed use limited to: Rajamangala Univ of Technology Thanyaburi provided by UniNet. Downloaded on March 03,2025 at 01:45:02 UTC from IEEE Xplore. Restrictions apply.
363