2023 IEEE Latin American Electron Devices Conference (LAEDC) Puebla, México, July 3-5, 2023

IoT-Based Neonatal Incubator Monitoring System

Alisson Waleska Martínez Fernanda de Lourdes Cáceres Kevin Fabricio Martínez
Engineering Faculty Engineering Faculty Engineering Faculty
Universidad Tecnológica Universidad Tecnológica Universidad Tecnológica
Centroamericana (UNITEC) Centroamericana (UNITEC) Centroamericana (UNITEC)
2023 IEEE Latin American Electron Devices Conference (LAEDC) | 979-8-3503-1190-7/23/$31.00 ©2023 IEEE | DOI: 10.1109/LAEDC58183.2023.10209121

Tegucigalpa, Honduras Tegucigalpa, Honduras Tegucigalpa, Honduras

aliwalem@unitec.edu fernanda_lcl@unitec.edu kevin.cruz@unitec.edu

Abstract—Neonatal incubators are essential in neonatology ranges defined by both the manufacturer and the institution
rooms and, as they are life support equipment, they require a that owns the equipment [7].
maintenance frequency of at least every six months to guarantee
their performance and safety. To analyze them, several devices There are neonatal incubator analyzers from different man-
exist designed to measure parameters such as temperature, ufacturers and different models. However, one of the most
humidity, and noise level inside the baby´s chamber. A complete widely used analyzers worldwide is the Fluke INCU II. It is
analysis consists of 9 tests with a duration of 5 hours and 56 easy to use and gives reliable results ensuring the safety of
minutes overall. During this time, the workflow of the incubator babies [7].
is interrupted. The objective of this research is to develop an
IoT-based prototype capable of monitoring the parameters inside The UNE-EN 60601-2-19 standard describes the essential
the chamber; without stopping the operation of the equipment safety requirements for the proper operation of neonatal incu-
and that provides the security of not affecting the health of bators, ensuring the health of the newborn and the operator.
the premature baby. The IoT architecture was designed and the This standard contains the mechanical and electrical safety
safety criteria for each component used to build the prototype
were determined. Experimentation was used to perform tests in guidelines, as well as regulations related to protection against
simulated environments and in a real environment. The results excessive temperatures and alarm systems [8].
measured by the prototype were compared with those obtained The Fluke analyzer allows you to verify all international
by the certified measurement instrument Fluke INCU II. The standards: IEC 60601-2-19, IEC 60601-2-20 and IEC 60601-2-
values were compared using non-parametric statistic analysis to
determine points for improvement. 21. This device allows to check incubators and radiant heaters.
Index Terms—biomedical engineering, IoT, monitoring system, It performs 9 different tests [7].
neonatal incubators, prototype It is important to mention that each test has a different
duration time. The analyzer needs 5 hours and 56 minutes to
I. I NTRODUCTION run all tests. It is not possible to analyze an incubator without
stopping it’s use and altering the workflow within a hospital.
Neonatal incubators are the medical devices that create an This research project proposes the development of an initial
ideal environment for premature babies where temperature, prototype capable of performing certain tests of a neonatal
humidity, noise, and oxygen flow are regulated [1]. This incubator analyzer that does not interrupt workflow and is safe
devices provide a protected environment for low birth weight for the baby. This device allows real-time monitoring of certain
infants [2]. These are used to increase the chances of survival environmental parameters created by the incubator.
of premature babies in the extrauterine environment [3]. It This paper is divided in chapters. Chapter II shows the
allows the baby’s body reduce the heat loss [4]. state of the art. Chapter III the methodology used for the
A neonatal incubator has a chamber to place the baby development of the prototype, IV shows the results and finally
covered by a transparent canopy. Several access ports allow chapter V the conclusions.
the medical team access to the baby for feeding, examination
and treatment. The baby is laid on a mattress inside located
on top of a fan and heater. These are responsible for providing II. S TATE OF THE A RT
a forced circulation of hot air within the compartment [5].
The malfunction of neonatal incubators can cause serious Multiple medical devices, sensors, diagnostic and imaging
injuries and even death of premature babies. The most frequent devices can be seen as smart objects. IoT-based healthcare
factors that cause these events are: the lack of inspection, the services are expected to reduce costs, increase quality of life
lack of an analysis protocol, errors in the production of said and enrich the user experience [9]. The IoT healthcare network
equipment and environmental noise [4] [6]. facilitates the transmission and reception of medical data, and
An incubator analyzer measures parameters such as temper- allowing the use of communications [15]. IoT in healthcare
ature, humidity, sound, and air circulation inside the incubator. has multiple applications like remote monitoring; this includes
It´s main objective is to verify that it is operating within the patients and clinical settings [11] [16].

979-8-3503-1190-7/23/$31.00 ©2023 IEEE

Authorized licensed use limited to: UNIVERSIDAD AUTONOMA DE BUCARAMANGA (UNAB). Downloaded on September 22,2023 at 19:23:53 UTC from IEEE Xplore. Restrictions apply.
 2023 IEEE Latin American Electron Devices Conference (LAEDC) Puebla, México, July 3-5, 2023

A. Temperature And Relative Humidity Verification Device standard was used to determine the safety requirements inside
Inside Infant Incubator Through IoT in Thailand neonatal incubators. The criteria taken into consideration for
The authors of reference [6] developed a device that mea- the development of the prototype are:
sures the temperature and humidity inside the chamber of • The device casing must be able to withstand temperatures

the neonatal incubator remotely through IoT. 5 sensors are of at least 38°C.
positioned on the baby’s mattress that transmit the data to • The temperature sensor must have an accuracy of ± 0.8

Thing Speak platform via Wi-Fi. °C.

• The humidity sensor must have an accuracy of ± 10
B. Autonomous Incubation System for the Monitoring of Pre- • There must be a microphone amplifier module to measure
mature Infants in Romania the noise inside the incubator room.
The authors in [17] designed an incubation system in • The device must be located so that it is difficult for the
which temperature and humidity are controlled autonomously. baby to touch some of its surfaces.
Different software and platforms were used, which are: IBM A 38-pin ESP32 was used as a microcontroller, due to
Cloud, the IBM Watson IoT platform and MIT App Inventor. the ease with which it can transfer data via Wi-Fi. It can
The system collects information about the infant, sets the operate normally in a temperature range of -40°C to 85°C,
optimal conditions based on that, and displays the LCD screen. ideal for use inside the neonatal incubator compartment. It is
also characterized by it´s ability to operate continuously for
C. Emergency Alert System for Neonatal Unit using IoT
long periods of time, their small size and the possibility of
In reference [18] the author created a system that can mon- connecting different sensors and actuators that work together.
itor several incubators to solve the problem of not receiving An LCD screen shows the registered values, making it easier
early alerts to treat premature babies. If the system detects an for the user to read them. A 20x4 LCD screen was used with
alteration, an alert will be sent to the medical professionals the purpose of displaying the data in an orderly manner. The
through the Blynk IoT software application. I2C module was used to connect with the ESP32 with only 4
pins. The screen can operate normally in temperatures from 0
III. M ETHODOLOGY
to 50°C.
The research was developed with a mixed approach. Infor- A DHT22 sensor was used to measure the temperature and
mation from standards was used to select the safest compo- relative humidity (RH) inside the chamber. It’s accuracy of
nents for the prototype. Operational tests were done both in ±0.8°C and <10% RH matches the standard’s requirements.
a controlled environment and in a real one, the parameters For the noise leve detection the MAX4466 microphone
measured by the prototype and an incubator analyzer were module with amplifier was selected, which is capable of de-
recorded. They were then compared to determine the effec- tecting sounds and can work with the ESP32 microcontroller.
tiveness of the prototype. A copper plated Bakelite PCB on one side was used to build
According to [12] IoT architecture for medical applications the electronic circuit. A Risk Analysis Matrix was done for the
follow the three-tier framework that includes application, material selection of the device’s casing. PMMA acrylic was
network and perception layers. Fig. 1 shows the architecture used for the prototype casing because it supports temperatures
proposed as a baseline for design. The device´s sensors of up to 75°C and has good stability to ultraviolet rays, ideal
connect to a microcontroller with a Wi-Fi connection. for situations in which the baby is receiving photo-therapy.
It was decided not to use any type of acrylic glue because
when exposed to high temperatures it could emit toxic gases,
so screws were used to join the casing.
The interior of the device was covered with Expanded
Polystyrene (EPS) due to its ability to insulate heat and it
supports temperatures of up to 85°C withput emitting toxic
gases [13]. This material does not release harmful gases when
exposed to high temperatures and is one of the most resistant
to the adverse effects of moisture [14].
A 10000mAh power bank was used as an external power
supply for the microcontroller. It was decided to use a long
Fig. 1. Proposed IoT architecture micro-USB to USB cable to locate the power bank outside
the incubator and avoid exposing it to high temperatures. The
battery allows continuous operation for more than 10 hours.
IV. R ESULTS AND D ISCUSSION
Tests were carried out on the components operating indi-
A. A. Criteria for the Prototype´s Design vidually with the microcontroller and then tests on all the
From information obtained in bibliographic databases, the components working together. Fritzing software was used to
importance of monitoring temperature, relative humidity and perform the schematic design and the PCB diagram as seen
noise was determined. The UNE-EN 60601-2-19 /A1:2016 in Fig. 2.

2
Authorized licensed use limited to: UNIVERSIDAD AUTONOMA DE BUCARAMANGA (UNAB). Downloaded on September 22,2023 at 19:23:53 UTC from IEEE Xplore. Restrictions apply.
 2023 IEEE Latin American Electron Devices Conference (LAEDC) Puebla, México, July 3-5, 2023

ESP32 connects to a Wi-Fi signal and transfers the measured

data to the UBIDOTS IoT platform linked with a dashboard
token number. This platform receives the information from the
sensors and displays them on a dashboard. The last block con-
sists of the display, any electronic device with internet access
and that has the display link can monitor data transferred to
UBIDOTS.

Fig. 2. Circuit schematics on Fritzing

B. Prototype´s Case
The size of the device were designed so that does not
interfere with the access to the chamber. The casing has four
suction cups on its front face to firmly hold the device in place
and prevent it from falling.
Fig. 3 shows an isometric view of the prototype casing
in SolidWorks with it´s size in cm. A laser cutter was used
to make the acrylic cuts precisely. The size of the casing is
enough to keep all components safely inside and small to be Fig. 5. Prototype´s IoT architecture
placed inside the incubator, as seen in Fig. 4.
The prototype design provides the facility to open its
casing for maintenance in case any component fails. It also D. Prototype testing in controlled environment
provides the opportunity of cleaning the device to prevent Pilot tests of the prototype were carried out in a controlled
cross-contamination. environment, for which the control booths of the Industrial
Engineering laboratory of UNITEC, Tegucigalpa were used.
These cabins have sound, high temperature, low temperature,
and humidity simulators.
The prototype’s data was compared to a hygrometer and
decibel meter. The prototype was fixed to the wall of the
controlled cabin to see if the suction cups can keep it in place.
It was observed that data was transferred simultaneously to
UBIDOTS.
E. Prototype testing in real environment
Fig. 3. Isometric view of casing
Functional tests of the prototype were carried out in a
neonatal incubator at Hospital María in Tegucigalpa. A Fluke
INCU II was used simultaneously to compare the data. Fig. 6
shows both devices measuring the 3 parameters.

Fig. 4. Simulation inside incubator

Fig. 6. Performing test with Fluke INCU II

C. Final IoT architecture
Fig. 5 shows the final architecture, which includes every The tests lasted for 6 hours, the incubator was turned on
component included in the prototype after testing. The DHT22 at a control temperature of 36°C and humidity of 60% RH.
temperature and humidity sensor are connected via GPIO25 to The Fluke analyzer was placed inside the incubator enclosure,
the ESP32, and the MAX4466 is connected via GPIO39. general test was selected as the test type.
The next block is made up of the LCD screen that displays The values were manually recorded every five minutes for
the data when connected to the ESP32. Simultaneously, the the first hour, then every 10 minutes until the end of the second

3
Authorized licensed use limited to: UNIVERSIDAD AUTONOMA DE BUCARAMANGA (UNAB). Downloaded on September 22,2023 at 19:23:53 UTC from IEEE Xplore. Restrictions apply.
 2023 IEEE Latin American Electron Devices Conference (LAEDC) Puebla, México, July 3-5, 2023

hour and every hour until completing 6 hours. The results for Temperature measurements are significantly similar between
the temperature sensor are shown in Table I with the difference the prototype and the Fluke analyzer, which indicates this
between both devices. The same was done for the humidity parameter shows the real temperature inside the incubator and
and noise levels. can be used as an analyzer.
For the humidity and noise level was not the case; mea-
TABLE I surements are significantly different, which indicates those
T EMPERATURE READINGS IN 6 HOURS
parameters are not correctly displayed by the prototype. This
Time(min) Prototype(ºC) Fluke(ºC) Difference(ºC)
error could be solved using a correction factor on the code o
selecting more accurate sensors.
0 29.7 32.1 -2.4
5 30 32 -2 The next step for this research is improve the code to be able
10 31.2 32.1 -0.9 to emulate more tests performed by a certified analyzer.This
15 31.4 32.2 -0.8 prototype is the first step to develop an incubator analyzer that
20 31.6 32.2 -0.6
25 31.8 32.28 -0.48 can perform each test indicated by the standards.
30 31.9 32.3 -0.4
35 31.9 32.2 -0.3 ACKNOWLEDGMENT
40 32 32.3 -0.3
45 32.2 32.37 -0.17 Thanks to Hospital María de Especialidades Pediatricas, for
50 32.2 32.35 -0.15
60 33.6 33 0.6 its openness and willingness to perform the last tests for the
70 33.6 33 0.6 prototype. Specially, to Engineer Gracia Girón, for her time
80 33.4 32.71 0.69 and willingness to collaborate.
90 33.3 32.57 0.73
100 33.1 32.7 0.4 This work was supported by the Research Fund of Univer-
110 32.7 32.7 0 sidad Tecnológica Centroamericana (UNITEC), Honduras.
120 32.7 32.8 -0.1
180 33.3 32.81 0.49
240 33 32.75 0.25 R EFERENCES
300 33 32.86 0.14
360 33 32.82 0.18 [1] P. Tiam Kapen, M. Youssoufa, M. Foutse, J. Dongmeza Koudjou,
y F. de P. Mkankam Kamga, “A multi-function neonatal incuba-
tor for low-income countries: Implementation and ab initio social
impact”, Med. Eng. Phys., vol. 77, pp. 114–117, mar. 2020, doi:
To compare the data from both devices, SPSS was used 10.1016/j.medengphy.2019.10.021.
to determine if there is a significant difference between the [2] D. B. Zimmer, A. A. P. Inks, N. Clark, y C. Sendi, “Design, Control, and
3 parameters. First, a normality test using the Shapiro-Wilk Simulation of a Neonatal Incubator”, en 2020 42nd Annual International
Conference of the IEEE Engineering in Medicine and Biology Society
method was carried out due to the sample size being less than (EMBC), 2020, pp. 6018–6023. doi: 10.1109/embc44109.2020.9175407.
50. The results show that non of the three parameters have a [3] M. Koli, P. Ladge, B. Prasad, R. Boria, y N. J. Balur, “Intelligent Baby
normal behavior and non-parametric tests were needed. Incubator”, en 2018 Second International Conference on Electronics,
Communication and Aerospace Technology (ICECA), mar. 2018, pp.
A non-parametric sign test was carried out to determine if 1036–1042. doi: 10.1109/ICECA.2018.8474763.
the measurements of the prototype were significantly different [4] A. Ili Flores, H. J. Konno, A. M. Massafra, y L. Schiaffino, “Simultane-
from the Fluke. According to the results from SPSS the ous Humidity and Temperature Fuzzy Logic Control in Neonatal Incuba-
tors”, en 2018 Argentine Conference on Automatic Control (AADECA),
significance between the temperature readings is 0.664, which nov. 2018, pp. 1–6. doi: 10.23919/AADECA.2018.8577290.
is greater than 0.05, which means the data is significantly [5] Draeger, “Neonatal Incubators–NICU Incubators for Premature Babies
similar. For humidity and noise this was not the case, being 0 Draeger”, 2021. https://www.draeger.com/en-us_ us/Hospital/Neonatal-
Care/Incubators-NICU.
the significance value; which means those two parameters are [6] J. Prinyakupt y K. Roongprasert, “Verification Device for Temperature
significantly different. This is verified showing that humidity and Relative Humidity Inside the Infant Incubator via IoT”, en 2019
values vary by approximately 5 to 6% and the noise values by 12th Biomedical Engineering International Conference (BMEiCON),
nov. 2019, pp. 1–6. doi: 10.1109/BMEiCON47515.2019.8990351
5 to 6 dB.
[7] A. Y. K. Chan, Biomedical Device Technology: Principles and Design.
Springfield, Illinois, 2008.
V. C ONCLUSIONS [8] Fluke, “INCU II Infant Incubator Tester Fluke Biomedical”. 2016.
[Online]. Available: https://www.flukebiomedical.com/products/
A prototype was developed capable of emulating the general biomedical-test-equipment/incubator-radiant-warmer-analyzers/
test of temperature sensing, accuracy of relative humidity and incu-ii-incubator-radiant-warmer-analyzer.
the sound level inside the baby´s chamber of an incubator. The [9] G. E. Valdés-de la Torre, M. Martina Luna, A. Braverman Bronstein,
J. Iglesias Leboreiro, y I. Bernárdez Zapata, “Medición comparativa
prototype was designed considering certain safety criteria of de la intensidad de ruido dentro y fuera de incubadoras cerradas”,
the UNE-EN 60601-2-19 /A1:2016 standard and is capable Perinatol. Reprod. Humana, vol. 32, núm. 2, pp. 65–69, jun. 2018, doi:
of monitoring parameters inside it´s compartment without 10.1016/j.rprh.2018.06.002.
[10] NE, “UNE-EN 60601-2-19:2009/A1:2016 (Ratificada) Equipos elec-
altering the workflow. The measured data is displayed on an troméd...”, 2016. Available: https://www.une.org/encuentra-tu-norma/
LCD screen and on the UBIDOTS digital platform, which busca-tu-norma/norma?c=N0057743.
allows monitoring the parameters in real time remotely. [11] P. P. Ray, D. Dash, y N. Kumar, “Sensors for internet of medical
things: State-of-the-art, security and privacy issues, challenges and future
The values measured by the prototype were compared with directions”, Comput. Commun., vol. 160, pp. 111–131, jul. 2020, doi:
those of the Fluke INCU II analyzer inside the chamber. 10.1016/j.comcom.2020.05.029.

4
Authorized licensed use limited to: UNIVERSIDAD AUTONOMA DE BUCARAMANGA (UNAB). Downloaded on September 22,2023 at 19:23:53 UTC from IEEE Xplore. Restrictions apply.
 2023 IEEE Latin American Electron Devices Conference (LAEDC) Puebla, México, July 3-5, 2023

[12] M. Aledhari, R. Razzak, B. Qolomany, A. Al-Fuqaha and F. Saeed,

"Biomedical IoT: Enabling Technologies, Architectural Elements, Chal-
lenges, and Future Directions," in IEEE Access, vol. 10, pp. 31306-
31339, 2022, doi: 10.1109/ACCESS.2022.3159235.
[13] S. M. R. Islam, D. Kwak, MD. H. Kabir, M. Hossain, y K.-S.
Kwak, “The Internet of Things for Health Care: A Comprehensive
Survey”, IEEE Access, vol. 3, pp. 678–708, 2015, doi: 10.1109/AC-
CESS.2015.2437951.
[14] IDAE, “Guía Técnica para la Rehabilitación de la Envolvente
Térmica de los Edificios”. 2007. [Online]. Available:
https://www.idae.es/uploads/documentos/documentos\_GUIA\
_TECNICA\_EPS\_Poliestireno\_Expandido\_v06\_972d8feb.pdf
[15] S. Bhattacharya y M. Pandey, “Significance of IoT in India’s E-
Medical Framework: A study”, en 2020 First International Conference
on Power, Control and Computing Technologies (ICPC2T), ene. 2020,
pp. 321–324. doi: 10.1109/ICPC2T48082.2020.9071513.
[16] M. M. Dhanvijay y S. C. Patil, “Internet of Things: A survey of enabling
technologies in healthcare and its applications”, Comput. Netw., vol.
153, pp. 113–131, abr. 2019, doi: 10.1016/j.comnet.2019.03.006.
[17] H. Alqaheri, S. Radha, J. Chatterjee, S. SHOORIYA, S. J., y N. SATISH,
“Toward an Autonomous Incubation System for Monitoring Premature
Infants”, Stud. Inform. Control, vol. 30, pp. 121–131, dic. 2021, doi:
10.24846/v30i4y202111.
[18] V. Govindaraj, A. Thiagarajan, S. Uthirapathi, A. S. Murugiah and
D. D. Jeevagan, "Emergency Alert System for Neonatal Unit using
IoT," 2021 3rd International Conference on Advances in Computing,
Communication Control and Networking (ICAC3N), Greater Noida,
India, 2021, pp. 700-703, doi: 10.1109/ICAC3N53548.2021.9725481.

5
Authorized licensed use limited to: UNIVERSIDAD AUTONOMA DE BUCARAMANGA (UNAB). Downloaded on September 22,2023 at 19:23:53 UTC from IEEE Xplore. Restrictions apply.