Volume: 21, Issue: 1

Page: 20-31
International Journal of Science and Business
2023
Journal homepage: ijsab.com/ijsb

AI Based Baby Nursing Robot

Tarun Debnath & Pallab Kanti Podder

Abstract
Childcare is a tough job for every parent. In a joint family, parents get support
from the grandparents, who spend most of their time with the baby. The
situation is quite different for those who live in urban. Parents, who remain
busy at day time in their job, become tired after their work, and baby caring
at that particular time becomes quite impossible for them. More specifically
the situation for job holder parents having twin babies cannot be imagined
easily. This situation can be solved partially by a baby nursing robot, which
can provide basic support needs for a baby. Already a lot of work has been IJSB
done before and researchers are working to get a perfect solution. This paper Accepted 22 January 2023
Published 27 January 2023
focuses on the AI-based baby nursing robot which is an automated baby DOI: 10.5281/zenodo.7574912
rocker capable of classifying baby cry to understand baby’s needs. After the
sound classification, it provides necessary functions for the baby such as baby
rocking, and baby entertainment functionality while in the cradle, and in case
of a hungry cry it will take the baby to its parents. In addition, the proposed
system can provide baby urination information through GSM to the parent's
mobile. Also, it can provide some necessary information such as Gas leakage,
Smoke detection, and Temperature, Humidity and CO level of the home
environment. The robot will raise an alarm immediately and take the baby to
the parents if it detects any unsafe situation detected by the sensors.

Keywords: AI, Baby, Robot, Childcare, Nursing, Baby cry, Cry classification.

About Author (s)

Tarun Debnath (corresponding author), Lecturer, Department of Information and Communication

Engineering, Pabna University of Science and Technology, Pabna, Bangladesh.
Pallab Kanti Podder, Associate Professor, Department of Information and Communication
Engineering, Pabna University of Science and Technology, Pabna, Bangladesh.

20
 IJSB Volume: 21, Issue: 1 Year: 2023 Page: 20-31

1. Introduction
Infants or toddlers need parents’ attention for every hour of each day, in this modern era,
parents are involved in firm exploitation, office ferment, and personal work. So, they cannot
provide enough time to take care of their children due to their involvement in the above
activities. Grandparents and caregivers are the only two options available for working parents.
With consideration of social media, some incidents like caregivers’ attacks on infants in a cruel
way are reported frequently. Because this reason the necessity to protect baby with minimal
human interference arises (Kavitha et al., 2019). A baby nursing robot with some artificial
intelligence can solve this problem of the service holder parents. Already a lot of work on this
task is still ongoing by the researchers. Such a comprehensive analysis was conducted between
the traditional techniques and autonomous robotic applications by (David et al., 2022). In
another study, the author discusses the ethics of robotic applications (Li & Meng, 2022). This
paper concludes that the robotic approach to baby caring is important these days but a lot of
research is necessary to do that. By addressing the importance of robotic care for parents, we
proposed a methodology for baby nursing through an automated baby nursing robot that has
Intelligence. To do so, at first the baby's cry is detected through a transducer shown in Figure
2. Through which the mechanical sound signal is converted to its electrical nature for further
processing. By analyzing the signal intensity or signal amplitude level the sound signal is
classified as Discomfort/ Tired/ Hungry/ Belly Pain/ Burping cry (Symon et al., 2017; Joshi &
Mehetre, 2017). From this research, it is found that each type of cry has a different amplitude
level that maintains a pattern. In this work, the amplitude level is mapped into its analog value
as well as the amplitude level, and generated analog value maintains a proportional
relationship between each other. If the classification is detected as discomfort/ tired cry then
the baby cradle robot will entertain the baby through rocking for one minute. After completing
the rocking function the robot will take the baby to its mother immediately by following the
predefine line (Pakdaman et al., 2010; Kalbande & Koche, 2018). Finally, after reaching the
bedside, the robot will restart its rocking functionality for four minutes. If the cry is determined
to be a Hungry cry, the robot will transport the infant to the mother for feeding. Similarly, if the
infant experiences stomach pain or require burping, the robot will promptly take the baby to
its mother utilizing the line follower. To ensure safety environment, we have developed a safety
alarm system. From this, we can observe temperature, humidity, Liquefied Petroleum Gas
(LPG) leakage, smoke, and Carbon Monoxide (CO) in the baby’s cradle robot environment
through sensors (Gupta et al., 2019). The DHT11 sensor was utilized to monitor humidity,
temperature, and CO, LPG, and smoke detected by the MQ-2 sensor in the room environment
(Nasaruddin et al., 2019). If the parameters show any unexpected observations or LPG/ smoke
is detected, an alarm will sound to alert the parents about their kid. To make the system usable
our last concern is the baby’s urine detection. A baby diaper will usually cure the problem, but
using one for an extended period of time can cause a rash on the baby's skin. So, a urine
detection and mobile call generation system is developed for detecting a baby’s urination
situation. If the associated circuitry detects urine on the specially designed urine mat then two
mobile calls will be generated successively to a predefined number with a 1-minute delay
interval. The proposed system is developed and implemented for monitoring and caring device
of the baby in real-time.

2. Proposed Methodology
The project is initiated to solve the baby nursing issue for a parent who remains busy in the
daytime in their office and wants primary support from an automated system for their baby
when they are at home. The proposed system will solve their problem. At first, we developed a
mechanical baby rocker which includes four motors with a hard rubber wheel for movement.
The device is capable to carry one baby. We created five functions for the baby those are

21
 IJSB Volume: 21, Issue: 1 Year: 2023 Page: 20-31

integrated with the robot. The entire process of our proposed method is illustrated in the
following flow diagram Figure 1.

Figure 1: Flow diagram of our proposed baby nursing robot

2.1 Proposed Architecture

The baby cry detection and control system, automated swing module, smoke, and gas detection
arrangement, urine detection scheme, and automated movement system are all part of the
cradle's architectural design. The following is a full description of the individual phases:

2.1.1 Cry Detection and Control System

Five things can make a baby cry. The infant may be hungry, it may have gas, needs the mother's
affection, feel discomfort with urination or other problem, and belly pain (Myakala et al., 2017).
The baby wet case is solved using GSM in a later section and the rest four types of cries can be
grouped mainly into two categories. One type of cry involves the mother's assistance, while
another can be handled by entertaining the infant, such as rocking the baby in the cradle. To
differentiate the cries, we played all the sounds and observed the analog output data from the
sound sensor. The observed data shows discomfort data provides the peak amplitude is lower
than the other types of cry. While the other cry shows peak amplitude is comparatively higher.
It provides a lower analog value pattern in the total range of 0 to 1023. The rest of the sound
provides comparatively high analog values in the same range. The change is programmed to
detect and classify the cries using a PIC16F877A microcontroller. The system analyzes the
baby's cry based on these parameters and takes appropriate action. So, in case of the mother’s
need the robot can understand the need and take the baby to the mother using the line follower
function if needed. This can be defined as the intelligence of the device. Similarly, if a discomfort
cry is detected the rocking function will be performed at first and after a certain interval for
example 1 minute the robot will take the baby to the parents using a black line drawn on a
white surface.

22
 IJSB Volume: 21, Issue: 1 Year: 2023 Page: 20-31

Figure 2: Sound detection Module

2.1.2 Automatic Swing Module

When the baby cry is recognized and predicted to be a discomfort cry, the robot will begin
swinging utilizing a high torque gear motor. The rocking function is performed using a
pic16f877a microcontroller and a 5-volt DC-operated relay. The relay holds power to energize
the motor coil for a particular time and then releases the required contact to de-energize the
motor coil. The rocking functionality can be visualized in Figure 3.

Figure 3: Baby swing functionality of the system

2.1.3 Line Follower Functionality

A PIC16F877A microcontroller and an MG-996R standard servo motor execute this operation.
Generally, servo motors are available in two types. These are continuous servo and standard
servo. The standard servo can rotate from 0 to 180 degrees only whereas the continuous servo
can rotate from 0 to 360 degrees in a continuous manner. As a result, we used four continuous
servo motors to create the robot structure. Moreover, the structure has a weight and will
increase its weight after having a baby in it. So, we have taken a metal gear modified Tower Pro
MG996R servo motor. Now, the kid’s room may have some furniture and other devices. That is
why the robot requires a path to travel or obstacle avoidance. We choose the line following
solution as a safe solution for a predefined path. As a path, a black line will be created on the
white surface of the room. The custom cradle is shown below.

23
 IJSB Volume: 21, Issue: 1 Year: 2023 Page: 20-31

Figure 4: Baby cradle frame

The system is capable to take the baby to its mother in case of any emergency or in case of the
mother’s attention. The embedded software will handle the control operation for line followers
as needed. After reaching the baby's mother the robot will play a buzzer to confirm its existence
and ends its entire operation.

2.1.4 Safety-Assurance Arrangement

The safety ensures arrangement is designed to measure Temperature, Humidity, CO Level, LPG
Gas leakage, & Smoke Detection in the environment air. This is accomplished by two sensors,
the MQ-2 and the DHT-11. The MQ-2 sensor shows the parameters in the Particles Per Million
(PPM) unit. The module includes a 20x4 LCD Display which displays all the above-mentioned
measurements through this display. If any observation is found such as gas or smoke detection
or CO level increase the buzzer module raises an alarm immediately to inform the parents
about the environment (Al-Haija et al., 2013). Another common phenomenon in Bangladesh is
that the temperature falls at late night and due to sleepiness conditions, we can understand the
situation very hardly. This problem is addressed for the baby and solved in the project. The
temperature variation in the room will be detected by the temperature sensor and analyzed by
the microcontroller seen in Figure 5. The alert will sound through a buzzer if there are any
temperature-related irregularities. The following Figure 5 shows the safety assurance
arrangement test for the system.

Figure 5: Smoke, Gas & CO detection system circuitry

2.1.5 Urine Detection and Call Generation

A very common baby cry is the discomfort cry, which can occur due to various reasons. Such a
reason is an urination problem. This can be also called a baby wet condition. This is usually
solved by a diaper. But most of the time, long time diaper usage may create a rash. Leaving a
baby in a wet condition may cause baby rash and no one wants to do that at all. So, the detection

24
 IJSB Volume: 21, Issue: 1 Year: 2023 Page: 20-31

of baby urine after urination, faster detection can be achieved through a mobile call and this
can be achieved by the GSM module. The system includes a customized urine sensor made of
copper tape, a comparator module, a microcontroller, and a SIM900A GSM Module (Mukherjee
et al., 2019). In dry conditions the resistance between the two copper conductors becomes high
and the sensor provides a high analog value in the 10-bit resolution which is about 800. If the
baby becomes wet, the resistivity decreases and the sensor provides an analog value of about
250. The comparator compares this value to a threshold value that can be adjusted. In this
project, we have set the value to 512. As a result, the comparator circuit provides logic 1 for
dry conditions and logic 0 for wet detection. An Arduino Uno microcontroller is used to realize
this logic from the comparator circuit and depending on the logic a mobile call will be generated
immediately for support by exploiting the necessary AT commands. In this way, the entire
system works together to solve the problems relating to babies’ take care. The customized
urine detection sensor is shown in Figure 6 and the call generation control section can be
shown in Figure 7.

Figure 6: Customized urine detection sensor

Figure 7: Urine detection circuit using Arduino Nano and Sim900 GSM Module

2.1.6 Automated Movement System

Our baby nursing robot uses line-following technology for its navigation. The line sensing
method is not a new or innovative way but the application of line follower in baby nursing can
be considered as an innovative way of baby nursing. A black surface absorbs the light and a
white surface reflects the light. Maximum reflection is achieved from a white surface and
minimum reflection is achieved from a black surface. So, to find the line, we will exploit this
quality of light. An Infrared Ray (IR) sensor or an LDR (Light-Dependent Resistor) can be used
to detect light. We chose the IR sensor because of its superior precision. Five IR sensors
arrangement are installed on the robot's Front side at center alignment to detect the line. The
robot is then positioned on the line so that it can use any sensor combination from all the
possible combinations. The possible combinations and corresponding actions are defined in
the PIC microcontroller and the particular action is taken against a particular sensor output
combination. The actions are to provide logic combinations to the L298 motor driver for the

25
 IJSB Volume: 21, Issue: 1 Year: 2023 Page: 20-31

servo-driving mechanism. To power up the motor driver we have used a separate 12volt dc
power supply with a common ground configuration. By using the aforementioned technique,
the robot is navigated by operating the wheels that are attached to the motors, and controlled.

2.2 Processing and Control Technology

To construct the entire system we have used two types of microcontrollers as the processing
and control technology. One is a PIC microcontroller from a microchip and another is Arduino.
Here, the PIC16F877A microcontroller is used as the main decision device. It is used for
controlling all of the system except the safety assurance device and urine detection. A
microcontroller (MCU for microcontroller unit) is a small computer on a single VLSI integrated
circuit (IC) chip. It contains a numerous processor, 368 bytes RAM, 256 bytes ROM, and
operates in 4 to 5.5 volts along with programmable input/output peripherals. (Bermoy &
Salvan, 2022). These microcontrollers are mainly used to design embedded applications. An
SoC (System on a Chip) may connect the external microcontroller chips as the motherboard
components, but a SoC usually integrates the advanced peripherals like a graphics processing
unit (GPU) and Wi-Fi interface controller as its internal microcontroller unit circuits (Podder
et al. 2021). The PIC16F877A Microcontroller Pin-Out diagram is given in Figure 8 (a).

Another microcontroller, Arduino Nano is a small, complete, and breadboard-friendly board,

based on the ATmega328P and was released in 2008 shown in Figure 8(b). These
microcontrollers are popular nowadays because of their vast libraries. These microcontrollers
operate at 16 MHz, having 2KB Static Ram, 32 KB Flash memory, and 30 header pins for
controlling i/o devices. The large range of operating voltage of this microcontroller is 5-20 volts
(Nano, 2018).

Figure 8: (a) Pin-Out diagram of PIC16F877A (b) Pin-Out diagram of Arduino Nano

3. Result and Discussion

The proposed system can detect a crying sound and it can classify the cry sound into three
major classes based on the signal amplitude or intensity level. In this section, we will discuss
the mechanism behind this classification. First, we have used a secondary LCD display shown
at the bottom of the Figure 10, from which a crying sound is played and the analog values are
displayed and noted after processing with the elimination of noise in the normal room
environment. These cries are taken from the dataset and then observed by our system. Each
sound was donated by the individual parents and its length is 7 to 8 seconds (Donateacry-
corpus, 2019). We collected the analog value at every 150ms and kept these values in an array

26
 IJSB Volume: 21, Issue: 1 Year: 2023 Page: 20-31

then the maximum and minimum analog values is noted for further processing. These analog
values can be observed in table-1 given below. From the table some decisions are made:
i) The maximum analog value of Belly Pain/ Burp starts from 700 and the minimum analog
value starts from 95 to 200 on average.
ii) The maximum analog value of Hungry Cry starts from 500 and the minimum analog
value starts from 125 to 180 on average.
iii) The maximum analog value of Discomfort/ Tired Cry does not cross 500 and the minimum
analog value starts from 80 to 125 on average.
Based on these observations the classification is done through programming and the system
can classify a crying sound as either a Belly Pain/ Burp cry, Hungry Cry, or Discomfort/ Tired
Cry.
Table 1: Audio sound parameters range and corresponding action required.
Minimum Maximum Sound
Sl. Audio file Action Required
Analog Value Analog Value Category
1. 69BDA5~1.mp3 116 762 Belly Pain Take to parents
2. D6CDA1~2.mp3 92 760 Belly Pain Take to parents
3. D6CDA1~3.mp3 152 763 Belly Pain Take to parents
4. D6CDA1~1.mp3 212 761 Belly Pain Take to parents
5. 643D64~1.mp3 167 419 Belly Pain Take to parents
6. 5AFC6A~1.mp3 117 758 Burp Take to parents
7. 7E4B9C~2.mp3 97 659 Burp Take to parents
8. AF3088~1.mp3 103 608 Burp Take to parents
9. 79FF40~1.mp3 209 760 Burp Take to parents
10. AEA8AE~1.mp3 267 760 Burp Take to parents
11. C249BC~2.mp3 168 758 Hungry Cry Take to parents
12. CDA99F~1.mp3 175 583 Hungry Cry Take to parents
13. D1E236~2.mp3 126 651 Hungry Cry Take to parents
14. D6CDA1~2.mp3 179 500 Hungry Cry Take to parents
15. 7A2222~3.mp3 175 680 Hungry Cry Take to parents
16. 7B0E16~1.mp3 81 356 Discomfort Try Entertain
17. 10A404~1.mp3 85 393 Discomfort Try Entertain
18. 64ACB3~2.mp3 121 642 Discomfort Try Entertain
19. D6CDA1~3.mp3 82 290 Discomfort Try Entertain
20. D1CB71~1.mp3 99 495 Discomfort Try Entertain
21. F258A8~1.mp3 77 300 Discomfort Try Entertain
22. 03ADDC~1.mp3 104 383 Tired Cry Try Entertain
23. 06C4CF~3.mp3 97 424 Tired Cry Try Entertain
24. 7A2222~1.mp3 76 374 Tired Cry Try Entertain
25. 79FF40~1.mp3 114 298 Tired Cry Try Entertain
26. B2AABC~1.mp3 83 307 Tired Cry Try Entertain

Our proposed system considers different environmental parameters for baby nursing. Firstly,
it can detect the urination of the baby. It also considers LPG leakage, CO level in the air, smoke,
humidity, and temperature of the room where the baby will be nursed. If any of the parameters
of LPG, CO gas, smoke, humidity, and temperature reach an unsafe level, the robot will sound
an alarm to alert the parents. This circuitry can be shown in the following Figure 9:

27
 IJSB Volume: 21, Issue: 1 Year: 2023 Page: 20-31

Figure 9: Ensuring safety by monitoring different environment parameters (LPG, CO gas, smoke,
humidity and temperature)

From the table-1 case no. 1 to 10 & 11 to 15 baby needs the parent’s support. In this situation,
the nursing robot detects belly pain/ burp and hungry cry respectively. In both cases, it follows
the line to reach the parents. It is safe because the embedded program ensures all motors are
in off condition at all 1 or 0 sensor logic. It will navigate only to the marked area. The detection
can be realized from Figure 10.

Figure 10: Detection of Belly Pain/ Burp cry after sound play through LED

Figure 11: Detection of Discomfort/ Tired cry after playing sound through LED

For 16-26 no. case Discomfort/ Tired mode is detected and the baby entertainment function is
played. After baby entertainment, the robot will take the baby to the parents by line follower
navigation. It never continuously moves outside of the marked area. However, as it was a
discomfort/ tired cry. After taking the baby to the parents it will restart the rocking function

28
 IJSB Volume: 21, Issue: 1 Year: 2023 Page: 20-31

for the next 4 minutes. This can be realized from figure 11 and the path that goes to the parents
can be visualized in Figure 12 (a) & (b).

Figure 12: (a) Path for traverse of Baby Nursing Robot (b) Line detection of the cradle.

The following Figure 13 demonstrates the entire baby nursing robot setup.

Figure 13: The proposed baby nursing robot.

4. Conclusion
The baby nursing robot can help working parents with basic support for handling babies in this
technology-driven age. We developed an automated baby rocker which is capable of classifying
the cry sound and informing the parents about the baby’s need. Moreover, it helps us to detect
baby urine, provides rocking functionality, and finally generates alert tone in any possible risk.
When the baby needs the mother’s love or it is hungry, the robot takes the baby to its mother.
All the above mentioned technologies have been developed, implemented, and tested in real-
time. By considering all the issues, the system can be termed as an example of AI and we can
say that the baby nursing robot is developed based on Artificial Intelligence. However, at this
moment, the proposed system cannot learn anything which can be solved in the near future.
We are working to implement more advanced features like entertainment based on the facial
expression of the baby. In the future, the proposed system can be further improved by
advanced frontier technologies like Machine Learning (ML) or Deep Learning.

29
 IJSB Volume: 21, Issue: 1 Year: 2023 Page: 20-31

Acknowledgment
This research was supported under the Innovation Grant in Information and Communication
Technology by the ICT Division, Posts and Telecommunications Ministry of Bangladesh.
Tracking number: 19IF12650.

References
Al-Haija, Q. A., Al-Qadeeb, H., & Al-Lwaimi, A. (2013). Case Study: Monitoring of AIR quality in
King Faisal University using a microcontroller and WSN. Procedia Computer Science, 21,
517-521.
DOI: https://doi.org/10.1016/j.procs.2013.09.072
Bermoy, E. L. R., & Salvan, V. V. P. (2022). Development of a Microcontroller-Based Water
Fountain Control System.
David, L., Popa, S. L., Barsan, M., Muresan, L., Ismaiel, A., Popa, L. C., ... & Dumitrascu, D. L.
(2022). Nursing procedures for advanced dementia: Traditional techniques versus
autonomous robotic applications. Experimental and Therapeutic Medicine, 23(2), 1-9.
DOI: https://doi.org/10.3892/etm.2021.11047
Donateacry-corpus. (2019, 3 6). Retrieved from Github: https://github.com/gveres/donateacry-
corpus
Gupta, H., Bhardwaj, D., Agrawal, H., Tikkiwal, V. A., & Kumar, A. (2019, February). An IoT based
air pollution monitoring system for smart cities. In 2019 IEEE International Conference on
Sustainable Energy Technologies and Systems (ICSETS) (pp. 173-177). IEEE.
DOI: https://doi.org/10.1109/ICSETS.2019.8744949
Joshi, M. P., & Mehetre, D. C. (2017, August). IoT based smart cradle system with an Android app
for baby monitoring. In 2017 International Conference on Computing, Communication,
Control and Automation (ICCUBEA) (pp. 1-4). IEEE.
DOI: https://doi.org/10.1109/ICCUBEA.2017.8463676
Kalbande, A. G., & Miss Shraddha, O. K. (2018). Line Follower Robot. Journal for Research, 4.
Kavitha, S., Neela, R. R., Sowndarya, M., & Harshitha, K. (2019, March). Analysis on IoT Based
Smart Cradle System with an Android Application for Baby Monitoring. In 2019 1st
International Conference on Advanced Technologies in Intelligent Control, Environment,
Computing & Communication Engineering (ICATIECE) (pp. 136-139). IEEE.
DOI: https://doi.org/10.1109/ICATIECE45860.2019.9063773
Li, K., & Meng, Z. (2022). Ethics of Robotics Applications. In International Conference on Cognitive
based Information Processing and Applications (CIPA 2021) (pp. 325-330). Springer,
Singapore.
DOI: https://doi.org/10.1007/978-981-16-5857-0_41
Mukherjee, S., Ghosh, A., & Sarkar, S. K. (2019, March). Arduino based Wireless Heart-rate
Monitoring system with Automatic SOS Message and/or Call facility using SIM900A GSM
Module. In 2019 International Conference on Vision Towards Emerging Trends in
Communication and Networking (ViTECoN) (pp. 1-5). IEEE.
DOI: https://doi.org/10.1109/ViTECoN.2019.8899504
Myakala, P. R., Nalumachu, R., Sharma, S., & Mittal, V. K. (2017, October). An intelligent system
for infant cry detection and information in real time. In 2017 Seventh International
Conference on Affective Computing and Intelligent Interaction Workshops and Demos
(ACIIW) (pp. 141-146). IEEE.
DOI: https://doi.org/10.1109/ACIIW.2017.8272604
Nano, A. (2018). Arduino Nano. A MOBICON Company.
Nasaruddin, A. N., Tee, B. T., Musthafah, M. T., & Jasman, M. E. S. M. (2019). Ambient data
analytic on indoor environment monitoring for office buildings in hot and humid climates.
Data in brief, 27, 104534.
DOI: https://doi.org/10.1016/j.dib.2019.104534

30
 IJSB Volume: 21, Issue: 1 Year: 2023 Page: 20-31

Pakdaman, M., Sanaatiyan, M. M., & Ghahroudi, M. R. (2010, February). A line follower robot from
design to implementation: Technical issues and problems. In 2010 The 2nd International
Conference on Computer and Automation Engineering (ICCAE) (Vol. 1, pp. 5-9). IEEE.
DOI: https://doi.org/10.1109/ICCAE.2010.5451881
Podder, P., Debnath, T., Faruk, M. O., & Islam, M. S. (2021). A Microcontroller Based Efficient
Scheduling System for Air Conditioner Focusing on Maximum Electricity Savings Using
PWM Concept. International Journal, 10(2).
DOI: https://doi.org/10.30534/ijatcse/2021/1001022021
Symon, A. F., Hassan, N., Rashid, H., Ahmed, I. U., & Reza, S. T. (2017, September). Design and
development of a smart baby monitoring system based on Raspberry Pi and Pi camera. In
2017 4th International Conference on Advances in Electrical Engineering (ICAEE) (pp. 117-
122). IEEE.
DOI: https://doi.org/10.1109/ICAEE.2017.8255338

Cite this article:

Tarun Debnath & Pallab Kanti Podder (2023). AI Based Baby Nursing Robot. International
Journal of Science and Business, 21(1), 20-31. doi: https://doi.org/ 10.5281/zenodo.7574912

Retrieved from http://ijsab.com/wp-content/uploads/2073.pdf

Published by

31