ok

1. 1
1. Introduction
Pregnancy is a challenging period for a mother, with a range of significant emotional, physical
and hormonal changes. One of the things that pregnant women often face is stress levels which can
increase during pregnancy. The psychological burden that may occur, whether related to future
uncertainty or changes in the body, can affect the well-being of pregnant women. Therefore, pregnant
women are a group that is prone to experiencing stress during pregnancy.
Stress during pregnancy can contribute to illness directly, through its physiological effects, or
indirectly, through maladaptive health behaviors (for example, smoking, poor eating habits or lack of
sleep) [1]. It is important to motivate people to adjust their behavior and life style and start using
appropriate stress coping strategies. In modern times, the various stresses in human life can make the
overall experience highly challenging, to the point where life itself becomes quite disconcerting.
Procreation is considered a very gratifying experience by the women but the other factors have made
it different than a simple biological experience. Pregnancy is a unique period in the life of a woman.
During pregnancy, almost every organ of the mother’s body has to work harder in order to meet the
demands of the developing fetus. Each maternal organ adapts in a different way and at a different
time [2]. Period of pregnancy brings various other problems associated with life, which are not merely
biological but social, cultural, economic, political, and psychological in their content. These problems
are surmounting and can be potentially stressful and excessive energy expenditure [3].
Not only stress, energy expenditure (EE) is also an aspect that needs to be considered during
pregnancy. The process of fetal formation, body adjustments, and normal daily tasks can increase the
level of EE in pregnant women. Therefore, adequate rest and good self-care are very necessary to
maintain the health of the mother and fetus. Regular health monitoring and open communication with
medical personnel will help identify risks and ensure that pregnant women receive appropriate care
to minimize potential risks to themselves and their fetus during this crucial period of pregnancy.
Data survey from a total of 30 states and New York City in 2009-2010, nearly three-quarters of
women reported they had experienced at least one stressful event in the 12 months prior to the delivery
of their child. Common stressful events include moving, serious illness, financial stress, death of a
loved one, among others as shown in Figure1.
Figure 1. Mothers Experiencing Stressful Event Survey Data
Women who experience high levels of stress during pregnancy have 25-60% higher risk for
preterm delivery, even after accounting for the effects of other established risk factors, compared to
women with low levels of stress [4]. Stress before and during pregnancy has been linked to low
birthweight babies independent of preterm delivery [5,6]. Increased maternal psychosocial stress is
associated with vascular disorders, such as hypertension and preeclampsia, which are major medical

2. 2
reasons for preterm delivery [4]. These conditions are most common for women who are African
American, older, or in first-time pregnancies [5]. Increased maternal psychosocial stress is associated
with a variety of unhealthy behaviors such as poor diet or nutrition and smoking, which are also risk
factors for preterm birth [4].
Stress and EE in pregnant women is associated with a significant impact on the health and
development of infants, increasing the risks of infant mortality, low birth weight, and premature birth
[7]. Acute stress, especially early in pregnancy, has been linked to an increased risk of premature
birth [7]. Meanwhile, chronic stress can lead to various complications, including premature birth, low
birth weight, hypertension, and developmental delays in infants [7,8]. Post-traumatic stress disorder
combined with severe depression has been associated with a fourfold increased risk of premature
birth, regardless of prescription drug effects [9].
Some studies have indicated a connection between maternal stress and EE during pregnancy
and congenital heart defects in infants [10]. Maternal loss during the prenatal period and 6 months
preconception is associated with an increased risk of overweight or obesity in their children, possibly
due to exposure to severe early-life stress [11]. Women with high stress levels and EE during
pregnancy are more likely to have children with autism-like characteristics, with this association
being most pronounced in children born to women with high stress levels during the first trimester
[12-13]. A study also found that an increase in fetal cortisol levels can impact specific brain functions,
leading to cognitive reaction time disturbances in adolescence [4].
The risk of stress and EE is one aspect that pregnant women need to be seriously aware of.
Excessive stress conditions can have a negative impact on the health of the mother and fetus,
including increasing the risk of high blood pressure and premature labor. Therefore, it is very
important for pregnant women to undergo regular examinations regarding the health of themselves
and the fetus they are carrying. Monitoring the health of pregnant women during the 9 months of
pregnancy is generally divided into three trimesters. It is necessary to carry out regular monitoring.
These routine check-ups not only help detect potential health problems early, but also provide an
opportunity to plan necessary interventions to reduce those risks.
The current health monitoring system is primarily designed for use within clinical settings, such
as clinics, hospitals, and counseling centers, and is characterized by its large size. Monitoring is
typically conducted while the patient is in a bed. However, these systems lack suitability for stress
monitoring. Consequently, there is a growing recognition of the significance of monitoring and
recording physiological parameters of patients outside traditional clinical environments, both in
research and in the broader fields of applied physiology and medicine. The envisioned solution
involves the development of a monitoring system that can operate beyond clinical settings, enabling
continuous monitoring of patients even after discharge and facilitating personalized health
monitoring.
A Few research activities have proposed possible solutions for monitoring stress in pregnant
women. [14] reveals those pregnant women still using watches as media. This is felt to be very
disturbing to the work activities of housewives. Patent of [15] number US10178965 makes a
monitoring device with the shape of a wristwatch. Pregnant women can be monitored their steps as
long as the tool is used. The invention by [16] number CN113057608A introduces a smart wristband
tailored for monitoring pregnant women and their environment. This wristband is interconnected with
smart terminals like cloud servers and mobile phones. Comprising a body worn on the wrist, the smart
bracelet incorporates a controller responsible for gathering and processing information about the
pregnant woman's condition and environmental factors. This enables a real-time assessment of the
pregnant woman's status and surroundings, enhancing monitoring capabilities. Smart bracelets for
monitoring pregnant women are convenient to carry, can help pregnant women and family members
to assess, can better know the real-time state of the body and environmental conditions of pregnant
women, and can treat and care for pregnant women more scientifically and reasonable. The Apple
Watch is one of the commercial innovations, a smart watch equipped with a sensor to read the EE
that comes out of its user.

3. 3
The subsequent research [17-27] focuses on the development of a device capable of measuring
Photoplethysmography (PPG), Electrocardiography (ECG), and Galvanic Skin Response (GSR)
signals as innovative methods to monitor stress levels in pregnant women. This device is designed to
provide more comprehensive information about physiological responses to stress during pregnancy.
PPG is employed to measure pulse rate and oxygen saturation, ECG provides information about heart
activity, while GSR records changes in skin electrical activity. The combination of these three
parameters is expected to offer a deeper understanding of the complex interaction between the
cardiovascular system and skin response to stress in pregnant women. This innovation holds
significant potential to enhance pregnancy monitoring, enabling earlier identification of health risks
and strengthening efforts to provide more targeted and effective care for pregnant women
experiencing stress.
The research conducted in references [28-32] employs a novel approach by utilizing sensors
for respiration, skin temperature, and acceleration to detect stress. These sensors offer a multifaceted
perspective on physiological and behavioral indicators associated with stress. Respiration sensors
enable the measurement of breathing patterns, while skin temperature sensors provide insights into
variations in thermal regulation, and acceleration sensors capture motion-related data. By integrating
data from these diverse sources, the research aims to enhance the accuracy and granularity of stress
detection, offering a more comprehensive understanding of how physiological and behavioral factors
interact in response to stressors. This innovative application of sensor technology holds promise for
advancing stress monitoring techniques, potentially leading to more effective stress management
strategies in various contexts, including the assessment of stress in pregnant women [33-40].
Previous research on stress monitoring have primarily addressed stress alone, neglecting its
correlation with EE during the physical activities of pregnant women. However, stress is also
influenced by EE during routine activities or hospital visits for pregnancy checkups. Recognizing this
connection is essential for a more comprehensive understanding of stress factors during pregnancy.
In this research, the latest solution to the problem of monitoring stress and energy expenditure in
pregnant women is the creation of wearable monitoring of stress status and EE based on IoT machine
learning which is named "MomiCare". “MomiCare” mainly consists of 6 (six) bio-signal sensors
(ECG, GSR, PPG, STS, RR, Acceleration), microcontroller, RF transmitter, power supply, and
smartphone, which communicate with each other via Bluetooth. To achieve a low-cost system, the
proposed system will be implemented using commercial off-the-shelf (COTS). The proposed system
architecture is shown in Figure 1.
Figure 2. The Proposed Wearable Design for Pregnant Stress monitoring system architecture.

4. 4
Machine learning algorithms can be used to monitor stress and EE over time by analyzing
various factors such as the real health potential of pregnant women. This algorithm can then
recommend solutions that optimize performance or serve as an early warning for pregnant women.
The bio signal sensor only produces signals, which can then be translated into monitoring stress and
EE conditions in pregnant women on four scales; high stress, medium stress, low stress and normal.
Machine learning models can predict stress conditions and EE by analyzing various parameters such
as heart rate, blood pressure, body temperature etc. This model learns from historical data to estimate
the potential for stress and EE.
At its core, the platform functions as an advanced monitoring system that continuously tracks
and analyzes stress and EE status in daily activities. By utilizing IoT sensors embedded in smart
wearables, real-time data regarding heart rate, blood pressure, body temperature can be collected.
These data points are then processed and interpreted using advanced machine learning algorithms,
allowing the system to accurately predict and assess stress conditions and EE.
The importance of this technology lies in its ability to provide proactive warnings and
notifications to pregnant women, families and doctors to prevent excessive stress and EE. By
predicting potential stress and EE in advance, this is to maintain the health of the mother and fetus in
a healthy condition.
The focus of this research is selecting the right sensor for monitoring, making tools, laboratory
testing, field testing, and data processing. The main focus of the first research period was to develop
6 (six) bio-signal sensors (ECG, GSR, PPG, STS, RR, Acceleration) to detect stress and energy
expenditure in pregnant women. Construction of the project will be carried out within three years.
1. Development of data acquisition and wearable system for stress and EE data measurement.
2. Development of machine learning for prediction and analysis stress and EE.
3. Development of monitoring systems for stress and EE in pregnant women.
2. Research methods, steps, and implementation progress
The research methods, implementation steps, and implementation progress of each research
period are described as follows:
2.1 Planning for the first-period research
The first-year research focuses on the development of an innovative smart wearable design to
detect stress levels in pregnant women using six different sensors: ECG, PPG, GSR, skin temperature,
respiration, and an accelerometer. The initial step involves the integration of technology, ensuring
that all six sensors can operate synergistically and provide accurate data. The research also pays
attention to the comfort aspect by designing a wearable bra that is not only effective in measurements
but also comfortable for pregnant women to wear throughout the day.
Following the design phase, the next step involves the creation of a prototype for the wearable
bra with the six sensors—ECG, PPG, GSR, skin temperature, respiration, and accelerometer. This
prototype development is crucial for translating the conceptual design into a tangible and functional
product. The integration of these sensors into the bra prototype aims to collect real-time data on
physiological parameters, providing a comprehensive understanding of the maternal stress levels. The
prototype will undergo testing, involving pregnant participants, to validate its functionality, comfort,
and efficacy in stress detection.
The development of data processing algorithms remains a critical focus during the prototype
phase, ensuring that the collected information is accurately interpreted. The research will continue to
emphasize data security and privacy, adhering to strict standards to safeguard the sensitive health-
related information. As the wearable bra prototype takes shape, collaboration with healthcare experts
will intensify, ensuring that the device aligns with both technological and medical standards. This
multidisciplinary approach aims to deliver a reliable and practical solution for monitoring stress in
pregnant women, ultimately contributing to maternal and fetus well-being.

5. 5
2.1.1 Data acquisition for stress detection in pregnant women
Acquiring data from smart wearables for monitoring stress and energy expenditure in pregnant
women involves deploying advanced technologies to gain real-time insights into their physiological
responses. These wearables, equipped with sensors and sophisticated monitoring capabilities, allow
for continuous and non-intrusive tracking of key metrics related to stress levels and energy
expenditure. The data acquisition process encompasses the careful configuration and distribution of
smart wearables among a cohort of pregnant women volunteers. These devices seamlessly integrate
into their daily routines, providing researchers with a wealth of information that aids in understanding
the dynamic interplay between maternal well-being, stress factors, and energy utilization during
pregnancy.
The collected data from the smart wearables contribute to a comprehensive analysis, enabling
researchers to identify patterns, trends, and fluctuations in stress and energy expenditure levels over
time. The non-obtrusive nature of these wearables allows for a more natural and continuous
monitoring experience for pregnant women, minimizing disruptions to their daily lives. The insights
derived from this data acquisition process not only deepen our understanding of the physiological
aspects of pregnancy but also pave the way for the development of personalized interventions and
support systems tailored to the unique needs of expectant mothers. Data acquisition for stress
detection in pregnant women shown in Figure 3.
Figure. 3. Data acquisition for stress detection in pregnant women
Figure 3 shows the data acquisition process starting with reading bio signals from the PPG,
ECG, GSR, respiration, acceleration and skin temperature sensors. The signal results are then
validated by comparing the sensor values with standard health devices. To test the sensor's function,
a connection was made to the internet via Bluetooth and WiFi to transfer data from monitoring stress
and energy in pregnant women. These steps aim to ensure the accuracy and reliability of the sensors
in recording various health parameters. With technological integration involving various types of
sensors and internet connectivity, this system provides the potential for advanced, real-time health
monitoring of pregnant women.
2.1.2 Specifications of microcontrollers and sensors used to detect stress in pregnant women
The selection of microcontroller and sensor specifications is based on the criteria of being
connected in one system, accurate, low energy and can be connected to IoT. Researchers have carried
out various experiments and publications in the last 1 year related to this research. The following are
the specifications of the microcontroller and sensors that will be used.

6. 6
In the context of developing smart wearables designed for the monitoring of stress and energy
expenditure in pregnant women, the meticulous selection of a microcontroller assumes paramount
significance. The decision-making process is contingent upon critical criteria, including energy
efficiency, compact form factor, Bluetooth functionality, WiFi capability, and seamless integration
with the six sensors mandated for comprehensive monitoring purposes. Each microcontroller variant
presents distinctive advantages, encompassing power conservation attributes, support for Bluetooth
Low Energy (BLE), and the seamless integration of WiFi functionalities. Optimal selection hinges
upon the project's nuanced requirements and the accessibility of robust development resources,
accompanied by comprehensive documentation support. Microcontroller specifications in this
research is shown in Table 1.
Table 1. Microcontroller Specifications
Parameters Value Remarks
Microcontroller
Supported Battery
UART
Operating Voltage
Power Consumption
Flash Memory
SRAM
3x 12
PCB Size
Weight
Radio Module
STM32H747XI dual Cortex®-
M7+M4 32bit low power
Arm®
Li-Po Single Cell, 3.7V,
700mAh Minimum
4 x Port (2 With Flow Control)
3.3 Volt
2.6 mA
2Mbyte +1RAM
8 MByte
A4 (SDA), A5 (SCL)
4.61 x 2.44 x 0.98 mm
0.8 g
Wifi+Bluetoth
Connect with IDE
Re-Charger
A/D
Low power
Low Consumption
Saving data
-
Connected
Small Size
Tiny
Low Energy
The selection of PPG sensor specifications in this research is oriented towards the main criteria
which include energy saving, compact size, high accuracy, and integration capabilities with smart
wearables. The focus of this research is on monitoring stress and energy expenditure in pregnant
women, so the operational sustainability of the PPG sensor is crucial. The selected PPG sensor must
combine energy efficiency to suit the needs of the wearable device, minimal physical size for user
comfort, high accuracy for data reliability, and the ability to be integrated synergistically with smart
wearables. By considering these aspects, the choice of PPG sensor specifications is critical in ensuring
the success and relevance of this research in understanding the dynamics of stress and energy use in
pregnant women. PPG Sensor Specifications in this research shown in Table 2.
Table 2. PPG Sensor Specifications
Parameters Value Remarks
Power 1.8-3.3 V Low power
Interface I2C 6-pin
Signal Sensing
Size
PIN
Infrared (IR) LED
and red LED
5.6 mm X 2.8 mm X 1.2 mm
VIN, SCL, SDA, INT,
IR-DRV, R-DRV
pulse oximetry (SpO2)
and heartbeat (HR)
Tiny
Connected
The selection of EGG sensor specifications in this research is focused on the main criteria which
include energy efficiency, small dimensions, high accuracy, and the ability to integrate with smart
wearables. The main focus of the research was monitoring stress levels and energy expenditure in
pregnant women. Therefore, the selected EGG sensor must be able to optimize energy usage to suit

7. 7
the characteristics of the device being worn, have a minimum physical size for user comfort, provide
a high level of accuracy to ensure the reliability of the data produced, and can be integrated
synergistically with smart devices wearables. Through these considerations, the choice of EGG sensor
specifications was critical in ensuring the success and relevance of this research in an in-depth
understanding of the dynamics of stress and energy use in pregnant women. ECG Sensor
Specifications in this research shown in Table 3.
Table 3. ECG Sensor Specifications
Parameters Value Remarks
Power 2.0 V-3.3 V Low power
Interface I2C 6-pin
Signal Sensing
Size
PIN
Biomedical Electrode
3.5 cm X 3 cm
GND, 3.3V, AN, LO-,
LO+, SDN
Electrocardiogram
Tiny
Connected
The selection of GSR sensor specifications in the context of this research prioritizes main
criteria such as energy efficiency, small dimensions, high level of accuracy, and the ability to be
integrated with smart wearables. This research focuses on monitoring stress levels and energy
expenditure in pregnant women, so the selected GSR sensor must be able to optimize energy use for
consistent use over a long period of time, have a minimum physical size for user comfort, provide
accurate measurement results for data reliability collected, and can be seamlessly integrated with
smart wearable devices. By considering these aspects, the choice of GSR sensor specifications is
crucial in ensuring the success and relevance of this research in describing the dynamics of stress and
energy expenditure in pregnant women. GSR Sensor Specifications in this research shown in Table
4.
Table 4. GSR Sensor Specifications
Parameters Value Remarks
Power 3.3 V-5 V Low power
Interface I2C 4-pin
Signal Sensing
Size
PIN
Resistance Conductivity
24 X 20 X 9.8 mm
GND, 5V, NC, A0
Electrical Conductance
Tiny
Connected
In this research, the selection of skin temperature sensor specifications is focused on the main
criteria involving small dimensions, a high level of accuracy, and the ability to be integrated with
smart wearables. The main focus of the research is monitoring stress levels and energy expenditure
in pregnant women, so the selected skin temperature sensor must have a minimum physical size for
user comfort, provide a high level of accuracy to ensure the reliability of the data collected, and be
able to be integrated smoothly with smart devices wearables. By considering these aspects, the choice
of skin temperature sensor specifications is a critical aspect in ensuring the success and relevance of
this research in describing the dynamics of stress and energy expenditure in pregnant women. Skin
Temperature Sensor Specifications in this research shown in Table 5.
Table 5. Skin Temperature Sensor Specifications
Parameters Value Remarks
Power 4 V- 30 V Low power
Interface I2C 3-pin
Signal Sensing Body Temperature -55°C – 150°C

8. 8
Size
PIN
24 X 20 X 9.8 mm
GND, VCC, A0
Tiny
Connected
In the context of this research, the selection of respiration sensor specifications focuses on
main criteria such as small dimensions, high level of accuracy, and the ability to be integrated with
smart wearables. This research emphasizes monitoring stress levels and energy expenditure in
pregnant women, so the selected respiratory sensor must have a minimum physical size for user
comfort, provide a high level of accuracy so that the data collected is reliable, and can be easily
integrated with smart wearable devices. Through these considerations, the selection of respiratory
sensor specifications is crucial in ensuring the success and relevance of this research in
comprehensively understanding the dynamics of stress and energy use in pregnant women.
Respiration Sensor Specifications in this research shown in Table 6.
Table 6. Respiration Sensor Specifications
Parameters Value Remarks
Power 1.1 V- 1.8 V Low power
Interface I2C 4-pin
Signal Sensing
Size
PIN
Respiration
4.9 X 6.3 mm
GND, VCC, SC, SDA
/minutes
Tiny
Connected
In this research, the emphasis on selecting accelerometer sensor specifications is focused on
the main criteria, namely small physical size, high level of accuracy, and the ability to be integrated
with a smart wearable. The main focus of the research was monitoring stress levels and energy
expenditure in pregnant women. Therefore, the selected accelerometer sensor must have a minimal
size to increase user comfort, provide a high level of accuracy to ensure the reliability of the collected
data, and be easily integrated with smart wearable devices. By considering these factors, the selection
of accelerometer sensor specifications is a critical aspect in ensuring the success and relevance of this
research in holistically evaluating the dynamics of stress and energy expenditure in pregnant women.
Accelerometer Sensor Specifications in this research shown in Table 6.
Table 7. Accelerometer Sensor Specifications
Parameters Value Remarks
Power 1.8 V- 3.6 V Low power
Interface I2C 3-pin
Signal Sensing
Size
PIN
Acceleration
23 X 1.2 mm
GND, VCC, Z
Tiny
Connected
2.1.3 Design of a wearable sensor for stress detection in pregnant women
The design of a bra-shaped wearable to detect stress in pregnant women could include the use
of various sensors including Photoplethysmogram (PPG), Galvanic Skin Response (GSR),
Electrocardiogram (ECG), skin temperature, respiration, and acceleration. In this design, PPG sensors
can be placed in the chest area to monitor heart rate, GSR sensors in the skin area to measure the
skin's response to stress, ECG sensors to monitor the electrical activity of the heart, skin temperature
sensors to record changes in body temperature associated with stress, respiratory sensors to measure
respiratory rate, and an accelerometer to monitor physical activity.
Signals from these sensors can be collected and processed by a data processing unit integrated in
the bra or an external device. The data obtained can be used to identify stress patterns and provide

9. 9
information to pregnant women or related medical professionals. The development of this wearable
must pay attention to user comfort, safety and privacy, and ensure that the design is safe for use during
pregnancy. In addition, collaboration with health experts and research ethics needs to be considered
in the development and testing of these tools.
Figure 4. Design of a smart wearable sensor for stress detection in pregnant women
Figure 4 above illustrates the smart wearable design developed in the context of this research.
The selected wearable form is a brassiere, a deliberate choice owing to its ubiquity as an accessory
among pregnant women. The decision to model the smart wearable as a brassiere is informed by the
comfort and unobtrusiveness associated with this garment, particularly during periods of heightened
activity for pregnant individuals. The strategic placement and positioning of six sensors—namely
PPG, ECG, GSR, respiration, acceleration, and skin temperature—have been determined based on a
comprehensive review of relevant literature and insights gained from prior experimental endeavors
conducted by the research team.
2.1.4 Bio signal wearable system for stress detection in pregnant women
The Bio Signal Wearable system with six sensors, namely electrocardiogram (ECG), galvanic
skin response (GSR), Photoplethysmogram (PPG), skin temperature sensor (STS), respiration rate
sensor (RTS), and acceleration sensor (ACC), is designed to provide comprehensive monitoring of
various physiological parameters. The ECG sensor can measure the electrical activity of the heart,
the GSR can provide information about skin stress levels, the PPG can monitor heart rate and heart
rate variability, the STS measures changes in skin temperature associated with the stress response,
the RTS can record respiratory rate, and the ACC monitors physical activity. The integration of these
six sensors on the wearable enables real-time monitoring that can provide a holistic picture of the
user's health and stress levels. Data generated by these systems can be analyzed to detect stress
patterns, provide better health insights, and support timely interventions. It is important to ensure
ergonomic design, user comfort, and data security and privacy in the development of these wearable
systems [41-50].
This system will utilize a tiny microcontroller as the main brain which is responsible for
collecting data from six different sensors which include physiological parameters of pregnant women,
namely electrocardiogram (ECG), galvanic skin response (GSR), Photoplethysmogram (PPG), skin
temperature sensor (STS) respiration rate sensor (RTS), and acceleration sensor (ACC). This

10. 1
0
microcontroller will be programmed to pre-process and store data locally before transmitting it via
Bluetooth and WiFi modules. Data successfully transferred via this wireless connection will be
continuously synchronized with iCloud, allowing easy and continuous access from various devices,
including smartphones or computers, so that information about pregnant women's stress levels can be
accessed by the user or even by health professionals who can provide necessary monitoring and
advice.
In designing these systems, it is important to ensure data security and privacy. Therefore, every
step in the data transfer process, including the use of Bluetooth and WiFi, must be encrypted to protect
sensitive information. Additionally, integration with cloud storage platforms such as iCloud must
comply with applicable data security and privacy standards, so users can feel confident that their
health data is managed securely and ethically.
Figure 5. Functional testing of smart wearables
The functional testing of a smart brassiere designed to monitor stress levels and energy
expenditure in pregnant women involves a comprehensive evaluation to ensure optimal performance
and clinical relevance. The testing encompasses aspects such as the accuracy of sensors in recording
skin temperature, respiration, and body acceleration, which are utilized for measuring physical
activity. Additionally, assessments are conducted on the device's ability to deliver real-time data to
monitoring platforms and its compatibility with mobile applications or data processing systems. User
comfort and safety aspects are also scrutinized to ensure that the brassiere design adheres to
ergonomic standards and does not cause discomfort for pregnant women. Throughout the testing,
attention is paid to the device's power consumption to ensure energy efficiency, as well as the validity
and reliability of the generated data. The overall functional testing is essential to ensure that the device
can provide accurate and valuable information in monitoring stress levels and energy expenditure in
pregnant women, while delivering a satisfactory user experience and meeting health and safety
standards.
2.2 Planning for the second-period research
The second-year research is centered on the development of machine learning algorithms to
process the data obtained from the wearable bra, aiming to detect stress levels in pregnant women
using six types of sensors: ECG (Electrocardiogram), PPG (Photoplethysmogram), GSR (Galvanic
Skin Response), skin temperature, respiration, and accelerometer (acc).During this phase, the focus
shifts to the application of machine learning techniques for the analysis of the extensive dataset
collected by the wearable bra's sensors. The goal is to train the machine learning model to recognize
patterns and correlations within the physiological parameters monitored, ultimately providing an
accurate and real-time assessment of the maternal stress levels. The integration of machine learning

11. 1
1
capabilities enhances the device's ability to adapt and improve its accuracy over time as it processes
more data.
The research will also involve refining and optimizing the algorithms to ensure efficient and
reliable stress detection. This iterative process may involve adjusting parameters, evaluating different
machine learning models, and enhancing the algorithm's sensitivity and specificity. Additionally, the
machine learning model will be trained to adapt to individual variations, considering that stress
manifestations can differ among pregnant women.
Collaboration with data scientists, machine learning experts, and healthcare professionals
becomes crucial in this phase. Their expertise will contribute to the development of robust algorithms
that can handle the complexity of physiological data and provide meaningful insights into the stress
levels experienced by pregnant women [51-55]. By the end of the second year, the aim is to have a
machine learning model that effectively processes the data from the wearable bra, offering a reliable
and personalized tool for monitoring and detecting stress in pregnant women. This advancement holds
great potential in contributing to maternal well-being and enhancing prenatal care through innovative
technological solutions. Machine Learning System shown in Figure 6.
Figure 6. Machine Learning System
Figure 6 represents the machine learning system meticulously crafted within the purview of this
research. Participants, comprising pregnant women volunteers, engage with smart wearables whose
functionality and calibration have undergone rigorous testing. Bio-signals acquired from an array of
sensors are transduced into data, subsequently subjected to machine learning processes executed on
a computational platform. This data undergoes the requisite phases of training, validation, and testing
within the machine learning framework. The resultant outcome of this iterative process manifests as
a robust model poised to analyze stress and energy expenditure. Finally, the culminating model is
deployed onto a compact microcontroller for practical implementation.

12. 1
2
2.2.1 Using Convolutional Neural Networks (CNN) to Detect Sensor Signal
The implementation of Convolutional Neural Networks (CNN) for monitoring stress levels in
pregnant women involves collecting data from various sensors such as electrocardiogram (ECG),
galvanic skin response (GSR), Photoplethysmogram (PPG), skin temperature sensor (STS),
respiration rate sensor (RTS), and acceleration sensor (ACC). After gathering the dataset, the next
step is pre-processing the data to remove noise and synchronize the timing of data from different
sensors. Subsequently, a CNN architecture is designed to extract important features from each sensor,
with specific convolutional layers to handle spatial data. Data from each sensor are combined to allow
the model to understand the relationships between sensors and generate a holistic representation of
the stress levels in pregnant women. After undergoing training, validation, and performance
evaluation, the implementation is carried out in the daily stress monitoring system, with continuous
maintenance and model updates to address changes in data characteristics or desired performance
improvements [56-71].
It is crucial to ensure that the implementation of this technology aligns with ethical norms and
healthcare privacy standards. Collaboration with healthcare professionals in the design process, data
collection, and implementation is key to ensuring that stress monitoring in pregnant women using
CNN is not only technically effective but also safe and compliant with applicable health standards.
CNN scheme is shown in Figure 7.
Figure 7. CNN Diagram for Stress and EE Model Development
The acquired raw data is divided into three dataset parts, data training, data validation, and data
testing. The size of the data training is larger compared to the other two to ensure the model have
enough data to spot trends and correlation inside the database, promoting the better results.
Thenceforth the dataset undergoes pre-processing step to identify and address missing values, noise,
and other inconsistencies prior to feeding it into the CNN algorithm. The CNN algorithm started with

13. 1
3
convolutional operation, where the cleaned dataset is filtered by a set of kennels who will convolve
across every single data inside the dataset. The output of this process subsequently fed through an
activation function referred to as an activation map. The activation map containing a summary of
features extracted from the input through the convolutional process. Applying additional convolution
layers will yield a better output, as more layers contribute to generating a more complex outcome,
therefore enhancing the overall result.
Following this procedure, the output will proceed to pooling layer to summarize the features from
an activation map. Using pooling layer helps to mitigate the effect of overfitting. The most common
types of pooling operation that are applicable is max pooling and average pooling. Finally, a classifier
is deployed to categorize the output dataset and generate the result.
2.2.2 Hyperparameters for Processing Data from Smart Wearables
Optimized Convolutional Neural Networks (CNNs) with a diverse set of hyperparameters
serve as a highly promising tool in generating data for the detection of stress levels and electrodermal
activity (EE) in pregnant women. Hyperparameters such as learning rate, the number of layers, and
the size of convolutional filters directly impact the network's ability to extract patterns of features
relevant to stress levels. Additionally, the inclusion of hyperparameters like batch size and dropout
rate can help mitigate overfitting in the model, ensuring robust generalization to pregnant individuals
who may exhibit complex variations in stress levels [61-65].
The importance of selecting appropriate hyperparameters extends to specific considerations
within the context of stress detection and EE in pregnant women. Factors such as dataset size,
hormonal variations during pregnancy, and individual sensitivity to stress necessitate careful
adjustments in hyperparameter settings. Implementing techniques like transfer learning or fine-tuning
on CNN architectures can enhance the model's capability to recognize unique patterns associated with
stress levels and electrodermal activity in pregnant women. This tailored approach ensures the CNN's
effectiveness in accurately capturing the nuanced physiological responses during pregnancy.
Figure 8. Hyperparameters for Processing Data from Smart Wearables
Figure 8 shows the hyperparameter settings on the Convolutional Neural Network (CNN) smart
wearable to monitor stress in pregnant women which is very significant. Hyperparameters such as
batch size, dropout rate, and number of layers can be adjusted to suit the needs of wearable smart
devices, such as smart bras, which are used to monitor stress levels in pregnant women. In addition,
the small physical size and energy efficiency of the accelerometer sensor, respiration sensor, and

14. 1
4
other sensors are integrated in the device to ensure comfort and reliability in monitoring. By carefully
aligning hyperparameter settings with the physical design and functionality of smart devices, this
smart wearable can provide a holistic and effective solution in understanding and managing stress
levels and energy expenditure in pregnant women.
2.2.3 Test the Machine Learning function
The functional testing of the machine learning system was conducted utilizing Convolutional
Neural Network (CNN) for the detection of stress levels and Energy Expenditure (EE) in pregnant
women. As an integral component of the research methodology, smart wearables, validated and
calibrated for functionality, were employed by pregnant volunteers. Bio-signals acquired from
diverse sensors were transformed into data, subsequently processed through a Convolutional Neural
Network to derive a model capable of discerning stress levels and EE. The testing process
encompassed stages of model training, validation, and testing, culminating in the deployment of the
final model onto a microcontroller for practical application in monitoring the health of pregnant
women. Test the Machine Learning function shown in Figure 9.
Figure 9. Test the Machine Learning function
The process of deploying machine learning with Convolutional Neural Network (CNN) models
onto tiny microcontrollers, such as Arduino, for stress and energy expenditure (EE) detection involves
selecting lightweight CNN architectures like MobileNet or SqueezeNet to address resource
limitations. Furthermore, model optimization through pruning and quantization can mitigate
computational burdens and memory requirements. Subsequently, the model needs to be converted
into a format compatible with the microcontroller for efficient implementation in detecting stress and
electroencephalogram activities on such constrained devices.
2.3 Planning for the third-period research
The research plan for the third period includes several main focuses which include testing
pregnant women, developing network configurations for smart wearable applications, as well

15. 1
5
as developing web and mobile-based monitoring applications for measuring and assessing stress and
EE in pregnant women. Here is a summary of each focus:
1. Testing of Voluntary Pregnant Women regarding Stress and EE Data:
The testing phase involving voluntary pregnant women focuses on the collection of data relating to
stress levels and energy expenditure (EE). This process entails the deployment of smart wearables to
conduct tests, allowing for the acquisition of relevant physiological and activity-related information.
The collected data includes metrics such as skin temperature, respiratory rate, accelerometer readings,
and potentially other relevant parameters from the integrated sensors. Following the testing phase, a
comprehensive analysis of the obtained results was conducted to gain valuable insights into the stress
levels and energy expenditure patterns exhibited by pregnant women. This analytical process is crucial
for understanding the dynamics of stress and energy utilization during pregnancy, ultimately
contributing to the refinement and optimization of the smart wearable technology designed for
monitoring these critical aspects of maternal health.
2. Network Configuration Design and Settings for Smart Wearable Applications:
The design and configuration of the network for smart wearable applications involve two key aspects.
Firstly, the network configuration is strategically designed to provide robust support for seamless
communication between smart wearable devices and monitoring applications. This includes
determining the architecture, protocols, and data transfer mechanisms that optimize the efficiency and
reliability of data transmission. Secondly, the preparation of the network infrastructure is undertaken,
ensuring the deployment of necessary hardware and software components to establish a secure and
responsive connection between the smart wearable devices and the monitoring applications. This
process may involve considerations such as the implementation of wireless communication protocols
(e.g., Bluetooth or Wi-Fi), data encryption for security, and the incorporation of cloud-based solutions
for data storage and accessibility. The meticulous design and settings of the network infrastructure are
pivotal to the overall success and functionality of smart wearable applications, facilitating real-time
monitoring and data retrieval for a comprehensive understanding of relevant health metrics.
3. Development of a Web-Based Monitoring Application for Real-time Measurement and Assessment of
Stress and EE of Pregnant Women:
The development of a web-based monitoring application for real-time measurement and assessment of
stress and energy expenditure (EE) in pregnant women involves the creation of a user-friendly and
accessible platform. The application is designed to facilitate seamless interaction and provide
instantaneous insights into stress levels and EE metrics. Through web-based technologies, users,
including pregnant women and healthcare professionals, can access the monitoring application directly
from their browsers. The application encompasses functionalities for real-time data collection,
processing, and visualization, offering a comprehensive and user-centric interface for measuring and
assessing stress and EE. This includes features such as graphical representations of stress trends, real-
time updates on EE levels, and potentially alert systems for critical thresholds. The web-based nature
of the application ensures flexibility, allowing users to engage with the monitoring process from
various devices with internet connectivity. The development of such a platform contributes
significantly to enhancing the accessibility and efficiency of stress and EE monitoring for pregnant
women, offering valuable insights for timely intervention and support.
4. Development of a Mobile-Based Monitoring Application for Real-time Measurement and Assessment
of Stress and EE of Pregnant Women:
The development of a mobile-based monitoring application for real-time measurement and assessment
of stress and energy expenditure (EE) in pregnant women involves the creation of a dedicated
application designed for use on mobile devices. This application is tailored to provide a user-friendly
interface, allowing pregnant women and healthcare professionals to monitor stress levels and EE in
real-time using their smartphones or tablets. The mobile app incorporates features for measuring and
scoring stress and EE, making it easily accessible and navigable through various mobile devices. The
real-time nature of the application ensures prompt updates and insights into stress dynamics and energy
utilization. This mobile-based solution enhances convenience, allowing users to engage with the
monitoring process anytime and anywhere, contributing to a more personalized and accessible
approach to maternal health monitoring.

16. 1
6
5. Mobile-based Monitoring Application Development Functionality Testing:
The functionality testing of the developed mobile-based monitoring application involves a rigorous
evaluation process to ensure its reliability and alignment with research objectives. This testing phase
systematically assesses each feature of the mobile application to verify its proper functioning. This
includes examining real-time stress and energy expenditure measurement capabilities, user interface
responsiveness, data accuracy, and the seamless integration of measurement and assessment
functionalities. The testing aims to identify and rectify any potential bugs, glitches, or inconsistencies
in the application, ensuring that all features operate as intended and meet the specific requirements
outlined in the research objectives. The comprehensive functionality testing is essential to guarantee
the application's effectiveness, accuracy, and usability, thereby enhancing its potential as a reliable
tool for real-time monitoring and assessment of stress and energy expenditure in pregnant women.
6. Hardware Setup and Configuration for Smart, Low Power in Laboratory Scale:
The process of setting up and configuring hardware for smart, low-power devices on a laboratory scale
involves several key steps. Firstly, the hardware configuration for smart wearable devices is
established, considering factors such as sensor integration, microcontroller selection, and power
management strategies to ensure efficiency. This includes optimizing the power consumption of the
devices while ensuring sufficient resources are available for accurate data collection. Subsequently,
the configured hardware undergoes laboratory scale testing to validate its performance and
functionality. This testing phase includes assessments of power usage efficiency, resource availability,
and the overall reliability of the hardware setup. Through this meticulous hardware setup and
configuration process, the aim is to create a robust and energy-efficient foundation for smart wearables,
ensuring their suitability for real-world applications, particularly in the context of monitoring stress
and energy expenditure in pregnant women.
This plan covers the holistic aspects of testing, application development, and hardware
configuration for this research. It is important to continuously monitor and transmit progress at
each stage so that research objectives can be achieved well. Field Test Diagram of a smart
wearable sensor for stress detection in pregnant women shown in Figure 10.
Figure 10 Field Test Diagram of a smart wearable sensor for stress detection in
pregnant women

17. 1
7
Following the development of a systematically integrated application connecting smart
wearables with the Internet of Things (IoT) shown in figure 9, a comprehensive evaluation was
conducted. The smart wearable, fashioned as a brassiere, underwent testing with a group of volunteer
participants, specifically 10 pregnant students hailing from Indonesia and pursuing their studies in
Taiwan. The experimental protocol involved the participants utilizing the smart wearable brassiere,
designed for optimal comfort during the active phases of pregnancy, over a period of three
consecutive days. Throughout this duration, the brassiere was worn during routine daily activities to
assess its performance and usability.
2.4 Expected difficulties and solutions
This research project composes of main works involve electronic design, programming,
measurement, mechanical design, computer network, and air conditioning setup. The research project
will face some potential difficulties during the research work's progress. The potential difficulties and
solution are described as follows:
This research project consists of main work including design, sensor design, programming,
measurement, mechanical design, stress measurement and energy expenditure. The research project
will encounter some potential difficulties during the progress of the research work. The potential
difficulties and their solutions are described as follows:
1. This research project is a multidisciplinary research field that has research members (hosts
and research assistants) from different educational and knowledge backgrounds. Sometimes,
it is not easy to build research chemistry among research members. Therefore, regular meeting
and discussion times will be developed to create a conducive research atmosphere where each
research member can enjoy and be comfortable sharing ideas, problem solutions, and research
progress.
2. Several parts of the research are needed related to expertise in the health sector because this
research aims to detect stress and energy expenditure (EE) from pregnant women to prevent
health problems in the fetus. Besides, the process of making wearables requires students from
the field of mechanical engineering. The solution is that this research involved 1 doctoral
program student and 1 NCUT master's program students in assisting the research process for
3 years.
3. Testing wearables on pregnant women requires volunteers, so an approach and collaboration
with the nearest health facility is needed to find suitable research subjects.
2.5 Usage of important instruments
Monitoring stress and energy expenditure (EE) in pregnant women is carried out to prevent the
mother from being too tired and causing disturbances to the fetus. In the first year of research, the
main supporting tools are needed in the form of 6 (six) bio signal sensors (ECG, GSR, PPG, STS,
RR, Acceleration). In the second year the focus is on sensor installation and data integration. a
computer is needed as a media as well as Wi-Fi and Bluetooth so that data interrogation can be carried
out optimally. The process of testing battery life will also be carried out, so that the type of battery
that is most suitable for the wearable that has been designed is needed. while in the third year, a
smartphone and real research subjects are needed. field data collection process and data processing.
3. Projects and results expected to be completed
3.3 The first-period research
A. Expected completion of the project
A. Expected project completion
1. Complete HW/SW (hardware and software) design and development. the use of suitable sensor
types for reading stress and EE in pregnant women.
2. Carry out the process of installing sensors on wearables that can monitor stress and EE in
pregnant women.
3. Get a battery type that is suitable for the wearable that has been designed.
4. produce wearables that are comfortable for pregnant women to use when on the move.

18. 1
8
5. use of suitable sensor types to read stress and EE in pregnant women.
6. Helping pregnant women/families to monitor excessive activities that will interfere with the
health of the mother and fetus
B. Expected results (outcome)
1. Wearable prototype for monitoring stress and EE
2. Publication (1 poster at a national conference/symposium, 1 patent application, 1 publication at
an international seminar, and 1-2 publications in an international journal SCI).
3. Monograph book
4. Increase the knowledge and ability of researchers in the design, development, calibration and
testing of bio signal sensors
5. Participate in special national academic contests and international invention fairs.
6. Increase the knowledge of researchers about applications related to the bio signal sensor platform
7. Interaction between industry and academia to strengthen the transfer of knowledge and
technology for researchers and students.
8. Understanding of research and development in the professional field will further enhance the
quality of teaching and research.
9. Improving the literacy atmosphere of research members through literature review, writing
technical reports, writing patent documents and writing published papers.
10. Encouraging 1 doctoral student and 1 master students to complete their research
3.4 The second-period research
A. Expected completion of the project
1. Complete HW/SW (hardware and software) design and development. the use of suitable
sensor types for reading stress and EE in pregnant women.
2. Carry out the process of installing sensors on wearables that can monitor stress and EE in
pregnant women.
3. Get a battery type that is suitable for the wearable that has been designed.
4. produce wearables that are comfortable for pregnant women to use when on the move.
5. use of suitable sensor types to read stress and EE in pregnant women.
6. Helping pregnant women/families to monitor excessive activities that will interfere with the
health of the mother and fetus.
B. Expected results (outcome)
1. A prototype of Wearable prototype for monitoring stress and EE
2. Publication (1 poster at a national conference/symposium, 1 patent application, 1 publication at
an international seminar, and 1-2 publications in an international journal SCI).
3. Monograph book
4. Increase the knowledge and ability of researchers in the design, development, calibration and
testing of bio signal sensors
5. Participate in special national academic contests and international invention fairs.
6. Increase the knowledge of researchers about applications related to the bio signal sensor platform
7. Interaction between industry and academia to strengthen the transfer of knowledge and
technology for researchers and students.
8. Understanding of research and development in the professional field will further enhance the
quality of teaching and research.
9. Improving the literacy atmosphere of research members through literature review, writing
technical reports, writing patent documents and writing published papers.
10. Encouraging 1 doctoral student and 1 master students to complete their research
3.5 The third-period research
A. Expected completion of the project

19. 1
9
1. Complete HW/SW (hardware and software) design and development. the use of suitable sensor
types for reading stress and EE in pregnant women.
2. Carry out the process of installing sensors on wearables that can monitor stress and EE in
pregnant women.
3. Get a battery type that is suitable for the wearable that has been designed.
4. produce wearables that are comfortable for pregnant women to use when on the move.
5. use of suitable sensor types to read stress and EE in pregnant women.
6. Helping pregnant women/families to monitor excessive activities that will interfere with the
health of the mother and fetus.
B. Expected results (outcome)
1. A prototype of Wearable prototype for monitoring stress and EE
2. Publication (1 poster at a national conference/symposium, 1 patent application, 1 publication at
an international seminar, and 1-2 publications in an international journal SCI).
3. Monograph book
4. Increase the knowledge and ability of researchers in the design, development, calibration and
testing of bio signal sensors
5. Participate in special national academic contests and international invention fairs.
6. Increase the knowledge of researchers about applications related to the bio signal sensor platform
7. Interaction between industry and academia to strengthen the transfer of knowledge and
technology for researchers and students.
8. Understanding of research and development in the professional field will further enhance the
quality of teaching and research.
9. Improving the literacy atmosphere of research members through literature review, writing
technical reports, writing patent documents and writing published papers.
10. Encouraging 1 doctoral student and 1 master students to complete their research.
4. References
1. Glanz, K., Schwartz, M., Stress, coping, and health behavior. Health behavior and health
education: Theory, research, and practice, pages 211-236, 2008.
2. Williams DJ., Physiology of healthy pregnancy, In: Warrell DA, CoxTM, Firth JD, editors.
Oxford Textbook of Medicine 4th ed, Oxford: Oxford University Press; pp. 383–385, 2003.
3. Han Hj., et al., Objective Stress Monitoring based on Wearable Sensors in Everyday: Journal of
Medical Engineering & Technology April 2020 DOI: 10.1080/03091902.2020.1759707.
4. Wadhwa, PD., et al., The contribution of maternal stress topreterm birth: issues and
considerations. ClinPerinatol, 38(3):351-84, 2011.
5. Cardwell, MS., Stress: Pregnancy considerations. ObstetGynecolSurv. 68(2):119-29, 2013.
6. Witt, WP., et al, Maternal stressful life events prior to conception and the impact on infant
birthweightin the United States. American J Public Health, 104(S1):S81-9, 2014.
7. Hills AP, Mokhtar N, and Byrne NM., Assessment of physical activity and energy expenditure:
an overview of objective measures: Frontiers in Nutrition. Vol 1 (5). 2014.
8. Pande, A., et al., Energy Expenditure Estimation with Smartphone Body Sensors: Conference:
Proceedings of the 8th International Conference on Body Area Network .2013.
9. Nagy LE and King JC. Energy expenditure of pregnant women at rest or walking self paced: The
American journal of clinical nutrition. Pp 369-376. 1983.
10. Pande A., et al., Using Smartphone Sensors for Improving Energy Expenditure Estimation: IEEE
Journal of Translational Engineering in Health and Medicine, Volume 3, 2015.
11. Pal, BS and Hepsida. D., Real time monitoring of daily calorie expenditure using smart phone:
International Journal of Advanced Research in Computer Science and Software Engineering.
Vol 3. (3). 2013.

20. 2
0
12. Most, J., et al., Energy intake requirements in pregnancy: Nutrients 2019, 11, 1812;
doi:10.3390/nu11081812
13. Pinheiro AC., et al., Energy expenditure: components and evaluation methods: Nutr Hosp.
2011;26(3):430-440.
14. Hobel CJ., Goldstein, A., Barrett, ES., Psychosocial stress and pregnancy
outcome.ClinObstetGynecol. 51(2):333-48, 2008.
15. Howell TA., et al., Activity Monitoring System for Pregnant Women: United States Patent.
Number US10.178.965B2, 2019.
16. Yang G. Smart Watch for Pregnant Mother: China Patent. Number N 113057608 A. 2021.
17. J. D. Runkle, M. M. Sugg, D. Boase, S. L. Galvin, and C. C. Coulson, "Use of wearable sensors
for pregnancy health and environmental monitoring: Descriptive findings from the perspective
of patients and providers," Digital Health, vol. 5, 2019.
18. R. Hiyama, M. Saito, Y. Nakanishi, Y. Hirose, and S. Arisumi, "BabyBumper:
protector/communication wearable device for pregnant women," Adjunct Proceedings of the
2015 ACM International Joint Conference on Pervasive and Ubiquitous Computing and
Proceedings of the 2015 ACM International Symposium on Wearable Computers, 2015.
19. J. Liu, B. Mosavati, A. V. Oleinikov, and E. Du, "Biosensors for Detection of Human Placental
Pathologies: A Review of Emerging Technologies and Current Trends," (in eng), Transl Res,
vol. 213, pp. 23-49, Nov 2019, doi: 10.1016/j.trsl.2019.05.002.
20. R. B. Patil and P. Krishnamoorthy, "Non-wearable sensor-based approach to monitor primary
health conditions," 2017 IEEE Region 10 Symposium (TENSYMP), pp. 1-4, 2017.
21. S. Nita, S. Bitam, and A. Mellouk, "A Body Area Network for Ubiquitous Driver Stress
Monitoring based on ECG Signal," 2019 International Conference on Wireless and Mobile
Computing, Networking and Communications (WiMob), pp. 1-6, 2019.
22. H. Niela-Vilén et al., "Pregnant women's daily patterns of well-being before and during the
COVID-19 pandemic in Finland: Longitudinal monitoring through smartwatch technology," (in
eng), PLoS One, vol. 16, no. 2, p. e0246494, 2021, doi: 10.1371/journal.pone.0246494.
23. A. Sano, S. Taylor, A. W. McHill, A. J. Phillips, L. K. Barger, E. Klerman, and R. Picard,
"Identifying Objective Physiological Markers and Modifiable Behaviors for Self-Reported Stress
and Mental Health Status Using Wearable Sensors and Mobile Phones: Observational Study,"
(in English), J Med Internet Res, Original Paper vol. 20, no. 6, p. e210, 2018, doi:
10.2196/jmir.9410.
24. C. N. Ginting, I. N. E. Lister, M. Silaen, E. Girsang, Y. Laila, M. Turnip, and A. Turnip, "Fetal
heart detection based wide area network technology with wireless sensor transmission," Journal
of Physics: Conference Series, vol. 1230, no. 1, p. 012037, 2019/07/01 2019, doi: 10.1088/1742-
6596/1230/1/012037.
25. C.-M. Wu, S.-C. Chen, and Y.-J. Chen, "A multiple bio-signal measurement analysis and
warning system for the long-term health care of severe disabled," Microsystem Technologies,
vol. 24, no. 1, pp. 155-163, 2018/01/01 2018, doi: 10.1007/s00542-016-3215-6.
26. C.-M. Wu, C. Y. Chuang, Y.-J. Chen, and S.-C. Chen, "A new estimate technology of non-
invasive continuous blood pressure measurement based on electrocardiograph," Advances in
Mechanical Engineering, vol. 8, no. 6, p. 1687814016653689, 2016/06/01 2016, doi:
10.1177/1687814016653689.
27. K. Dese, G. Ayana, and G. Lamesgin Simegn, "Low cost, non-invasive, and continuous vital
signs monitoring device for pregnant women in low resource settings (Lvital device)," (in eng),
HardwareX, vol. 11, p. e00276, Apr 2022, doi: 10.1016/j.ohx.2022.e00276.
28. Z. He, K. Wang, Z. Zhao, T. Zhang, Y. Li, and L. Wang, ‘A Wearable Flexible Acceleration
Sensor for Monitoring Human Motion’, Biosensors, vol. 12, no. 8, p. 620, Aug. 2022, doi:
10.3390/bios12080620.
29. J. Wojciechowski and E. Skrzetuska, ‘Creation and Analysis of a Respiratory Sensor Using the
Screen-Printing Method and the Arduino Platform’, Sensors, vol. 23, no. 4, p. 2315, Feb. 2023,
doi: 10.3390/s23042315.

21. 2
1
30. J. González-Buesa and M. L. Salvador, ‘An Arduino-based low-cost device for the measurement
of the respiration rates of fruits and vegetables’, Computers and Electronics in Agriculture, vol.
162, pp. 14–20, Jul. 2019, doi: 10.1016/j.compag.2019.03.029.
31. BUET/Department of CSE, Dhaka, Bangladesh, S. M. Farhad, M. R. Minar, and S. Majumder,
‘Measurement of Vital Signs with Non-invasive and Wireless Sensing Technologies and Health
Monitoring’, JAIT, pp. 187–193, 2017, doi: 10.12720/jait.8.3.187-193.
32. B. Sumathy, S. Kavimullai, S. Shushmithaa, and S. S. Anusha, ‘Wearable Non-invasive Health
Monitoring Device for Elderly using IOT’, IOP Conf. Ser.: Mater. Sci. Eng., vol. 1012, no. 1, p.
012011, Jan. 2021, doi: 10.1088/1757-899X/1012/1/012011.
33. Behrman, RE., StithButler, A., editors. Institute of Medicine, Committee on Understanding
Preterm Birth and Assuring Healthy Outcomes.
34. Yonkers, KA., et al, Pregnant women with posttraumatic stress disorder and risk of preterm birth.
JAMA Psychiatry. Published online June 11, 2014. doi:10.1001/jamapsychiatry.2014.558, 2014.
35. Zhu, JL., et al. Prenatal maternal bereavement and congenital heart defects in offspring: A
registry-based study. Pediatrics, 131(4):e1225-30, 2013.
36. Hohwu, L., et al, Severe maternal stress exposure due to bereavement before, during and after
pregnancy and risk of overweight and obesity in young adult men: a Danish National Cohort
Study. PLoSOne. 9(5):e97490, 2014.
37. Walder, DJ., et al, Prenatal maternal stress predicts autism traits in 61/2 year-old children: Project
ice storm.
38. Butte NF and King JC. Energy requirements during pregnancy and lactation: Public Health
Nutrition: 8(7A), 1010–1027.
39. Coussons-Read ME. Effects of prenatal stress on pregnancy and human development:
mechanisms and pathways: Obstetric Medicine 2013; 6: 52–57. DOI:
10.1177/1753495X12473751.
40. Harris A, Seckl J. Glucocorticoids, prenatal stress and the programming of disease: Horm Behav
2011;59:279 – 89.
41. Golawski K and Wojtya. Impact of Physical Activity of Pregnant Women on Obstetric
Outcomes: Int. J. Environ. Res. Public Health 2022, 19, 12541.
https://doi.org/10.3390/ijerph191912541.
42. Basjaruddin NC, Syahbarudin F, and Sutjiredjeki E. Measurement Device for Stress Level and
Vital Sign Based on Sensor Fusion: Healthcare Informatics Research, 2021.
https://doi.org/10.4258/hir.2021.27.1.11 .
43. Han Hj., et al., Objective Stress Monitoring based on Wearable Sensors in Everyday: Journal of
Medical Engineering & Technology April 2020 DOI: 10.1080/03091902.2020.1759707.
44. Jovanov E., et al., Stress Monitoring Using a Distributed Wireless Intelligent Sensor System:
IEEE Engineering in Medicine and Biology Magazine. 2003.
45. Pande A., et al., Using Smartphone Sensors for Improving Energy Expenditure Estimation: IEEE
Journal of Translational Engineering in Health and Medicine, Volume 3, 2015.
46. Altini M and Amft O. Energy expenditure estimation using wearable sensors: A new
methodology for activity-specific models: Conference Paper, 2012.
https://www.researchgate.net/publication/262222471
47. Vavrinsky E, et al., Application of Acceleration Sensors in Physiological Experiments: Journal
of Electrical Engineering, September 2014 DOI: 10.2478/jee-2014-0049.
48. Most J E, et al., Energy Intake Requirements in Pregnancy: sini International Journal of
Advanced Research in Computer Science and Software Engineering Energy 2019, 11, 1812;
doi:10.3390/nu11081812Nutr Hosp. 2011;26(3):430-440.
49. Choi J. Development and Evaluation of an Ambulatory Stress Monitor Based on Wearable
Sensors: IEEE Transactions on Information Technology in Biomedicine, VOL. 16, NO. 2, March
2012.
50. Chen J, Abbod M, and Shieh JS. Pain and Stress Detection Using Wearable Sensors and
Devices—A Review: Sensors 2021, 21, 1030. https://doi.org/10.3390/s21041030.

22. 2
2
51. Fen Z, et al., a cross-sectional study on anxiety and stress in pregnant women with chronic hBV
infection in the People’s republic of china. Journal Neuropsychiatric Disease and Treatment,
2015:11 2225–2232.
52. Chen CH, Yu YM, Hwang KK. [Psychological stressors perceived by pregnant women during
their third trimester]. Formosan Journal of Public Health. 1983;10(1):88Y98. Chinese.
53. Chang HS, Chen SY, and Chen CH. Predictors of Antenatal Psychosocial Stress in Taiwanese
Women. Journal of Nursing Research 2016, 24(3) p 193-200.
54. Hee J,H, Objective Stress Monitoring based on Wearable Sensors in Everyday Settings. Thesis.
University of California. 2019.
55. Gupta A, Bhandari S. ECG Noise Reduction By Different Filters - A Comparative Analysis.
International Journal of Research in Computer and Communication Technology. 2015;
4(7):424–431.
56. Deng Y, Wu Z, Chu CH, et al. Evaluating feature selection for stress identification. In: 2012
IEEE 13th International Conference on Information Reuse & Integration (IRI); IEEE; 2012. p.
584–591.
57. Hall MA. Correlation-based feature selection for machine learning; 1999.
58. Shaffer F, Ginsberg J. An overview of heart rate variability metrics and norms. Frontiers in
public health. 2017;5:258.
59. Das D, Datta S, Bhattacharjee T, et al. Eliminating individual bias to improve stress detection
from multimodal physiological data. In: 2018 40th Annual International Conference of the IEEE
Engineering in Medicine and Biology Society (EMBC); IEEE; 2018. p. 5753–5758.
60. Altman NS. An introduction to kernel and nearest-neighbor nonparametric regression. The
American Statistician. 1992;46(3):175–185.
61. Cortes C, Vapnik V. Support-vector networks. Machine learning. 1995;20(3):273–297.
62. Chang CC, Lin CJ. LIBSVM: A library for support vector machines. ACM Transactions on
Intelligent Systems and Technology. 2011;2:27:1–27:27. Software available at http://
www.csie.ntu.edu.tw/ cjlin/libsvm.
63. Hand DJ, Yu K. Idiot’s bayesnot so stupid after all? International statistical review. 2001;
69(3):385–398.
64. Blum A, Mitchell T. Combining labeled and unlabeled data with co-training. In: Proceedings of
the eleventh annual conference on Computational learning theory; ACM; 1998. p.92–100.
65. F Albinali et all. Using wearable activity type detection to improve physical activity energy
expenditure. In Ubicomp, Pages 311-320, 2010.
66. A.G. Bonomi et all. Improving assessment of daily energy expenditure by identifying types of
physical activity with a single accelerometer. Journal of Applied Physiology, 107(3):655-661,
2009.
67. E. Tapia. Using machine learning for real-time activity recognition and estimation of energy
expenditure. In PhD thesis, MIT, 2008.
68. M. Abitbol. Effect of posture and locomotion on energy expenditure. American Journal of
Physical Anthropology, 77(2):191–199, 1988.
69. David R. Bassett Jr, Alex R, and Stewart G. Calibration and Validation of Wearable Monitors.
Journal of the American College of Sports Medicine, Page: S32-S38. 2012.
70. Ann G et all. Harris–Benedict equation estimations of energy needs as compared to measured
24-h energy expenditure by indirect calorimetry in people with early to mid-stage Huntington’s
disease. Nutritional Neuroscience. Page: 213-218. Vol 11 No 5, 2008.
71. Aleksandra Zając, MD dan Mateusz Mucha. Harris-Benedict Calculator (Total Daily Energy
Expenditure). https://www.omnicalculator.com/health/bmr-harris-benedict-equation. 2023.