Stress Measuring Final Report April 20th, 2015
Stress Measuring Final Report
April 20th, 2015
This paper discusses all of the elements and information that went into designing a
Stress measuring application that is implemented in a wearable device. The problem is
clearly defined in a problem statement, and from this statement, a list of project objectives
and a literature survey are performed. Realistic constraints that guided the project and
Marketing Requirements that ensure the device is competitive in the wearable device
market are also created to guide the project. After this information is collected to guide the
project, the proposed designs are discussed. One of these is chosen and described further
in the design solution section. Based on this design, the integration of different parts, the
tests that are done to verify the design are discussed. Lastly, the budget and schedule are
included in this report.
Throughout this project, we accepted the help and guidance of many people
throughout the college of engineering. First, we would like to show our gratitude to our
Advisor Ziad Youssfi for giving us guidance, advice, and assisting in deciding what project
to work on. We would also like the thank all of the professors for giving constructive criticism
and counsel on all of our presentations and reports throughout the development of the
project and throughout the year.
Table of Contents
ACKNOWLEDGMENTS ............................................................................................... 2
LIST OF FIGURES........................................................................................................ 4
LIST OF TABLES.......................................................................................................... 5
PROBLEM STATEMENT.............................................................................................. 6
PROJECT OBJECTIVES.............................................................................................. 6
LITERATURE SURVEY (BACKGROUND)................................................................... 6
PROJECT REALISTIC CONSTRAINTS....................................................................... 8
MARKETING REQUIREMENTS………......................................................................... 8
ALTERNATIVE DESIGNS............................................................................................. 9
THE DESIGN SOLUTION............................................................................................. 10
DESIGN INTEGRATION…........................................................................................... 15
DESIGN VERIFICATION............................................................................................... 18
DESIGN BUDGET………………………………………………........................................ 20
DESIGN SCHEDULE……............................................................................................. 21
List of Figures
Level 0 Block Diagram
Level 1 Block Diagram
Level 2 Block Diagram
Heart Rate Variability Diagram
Physiological Stress equation
Android Application Screenshots
(Main Page, Survey Question, Survey Results)
ADS1293 Functional Block Diagram
SPI protocol from ADS1293
ECG Graph (Texas Instruments Software)
ECG Graph (Excel data from Arduino)
ECG Graph (Android Application)
Gantt Chart Graphical Representation
List of Tables
Alternative Solutions Decision Matrix
Stress Monitor Module Description
Psychological Survey Module Description
Irregular Heart Rate Monitor Module Description
Sensoria Smart Shirt Module Description
ADS1293 Module Description
Arduino Uno Module Description
Bluetooth Device Module Description
Heart Rate Variability Study Data
Device Unit Cost
Work Breakdown Structure
According to the IHS (International Handling Services) Wearable Technology Market
Assessment the current market for wearable technology is around 10 billion dollars. This is
expected to triple in the next five years . This study also says that there is a military
market driver for the development of smart clothes that transmit physiological parameters
including heart rate data. Variable heart rate, which is not measured by numerous devices
on the market, are important in determining how stressed a person is. Currently there are
no devices on the market that will measure stress using this physiological signal and are
small enough to wear around all day. The American Psychological Association conducted a
study on the impact of stress in 2012 and found that 20 percent of Americans reported that
their stress level is an 8, 9, or 10 on a 10-point scale . This indicates a large market for
this type of device because it is something that a lot of people struggle with.
The objective of this project is to design a rechargeable device that records
physiological signals and conduct a psychological survey to help individuals monitor their
stress level throughout the day. The physiological information is sent to a smartphone
application and displayed in a meaningful way. The displayed information can give
biofeedback to the person to lower their stress level. The physiological signal that we plan
to measure is irregular heart rate using a compact device with low power consumption to
allow battery operation through the day. This device is wearable throughout the day without
being a burden on the user.
Literature Survey (Background)
IHS (International Handling Services) has performed assessment on the Wearable
Technology Market. There were numerous important statistics from this report that could be
related to our problem. First, the market is projected to triple in the next five years. Both the
military and fitness market are in the process of developing smart clothes that can transmit
different physiological signals. The last important aspect of this report is that smart clothing
is a wearable technology that will create the most revenue in the future .
The final goal of our project is to display the user’s stress level. The American
Psychological Association put out a study on the impact of stress. This survey says that
64% of people believe that managing stress is important to them. Only 60% say that eating
healthy is important to them and 57% say that being physically fit is important to them. A
device that could help manage the stress of these individuals could attract a lot of interest
Research needed to be done on what defines smart clothes. Smart clothes are the
integration of technology and clothing that can
transmit data about the user’s vital signs. A few applications that this could be used for is to
supply doctors with real time data that could be used to help diagnostics. It could also link to
devices that would display when assistance was needed. Also, this could be used in the
fitness market to monitor physiological signals while working out or throughout the day .
We need to find ways to measure the variable heart rate. The American Heart
Association published an article on heart rate variability. The most common way to do this is
to take the heart rate for a specific period of time, either 24 hours or a short period of a few
minutes. Once this data is taken, the standard deviation of the time between the peak of
each beat is taken. A large standard deviation means that the user’s stress level is low and
a small standard deviation means that they user stress level is high. This is probably how
heart rate variability will be related to stress .
Once heart rate variability is measured, it is important to know if this value is out of
the normal range. A study was performed on the normal values of heart rate variability at
rest in a young, healthy, population. This study shows the mean and standard deviation
times between beats for different percentiles of this study which will be useful for
determining the normal range for our device. This is performed in the time and frequency
domain and by active individuals and individuals who train 6 hours a week .
This source is an article by the American College of Cardiology Foundation on Heart
Rate Variability and what this might mean. It also describes different ways to measure heart
rate variability. The most common ways are to use the time domain standard deviation and
frequency domain. It also goes into the future of measuring heart rate variability which is
based on nonlinear dynamics .
When looking up alternate ways to take cardiac pulse measurements we came
across a study that goes over non-contact, automated cardiac pulse measurements using
video imaging. They used a webcam to take a video of the user and used amplification and
analysis techniques to determine small changes in the user’s skin color. This can be used to
determine the pulse of the user .
As research was being performed, the new Apple Watch was announced which
contained a heart rate monitor that took the measurements from the wrist. From the
information on Apple’s website, it seems that they are taking this measurement the same
way as the newly released Samsung Galaxy Gear Watch. These devices have LEDs and a
camera/photodiode that can determine small changes in the skin color of the user that occur
on each pulse. From the reviews of the Samsung Galaxy Gear Fit, it seemed like this
approach was not very accurate and had not been perfected .
At MIT a project was done on amplification of the color of a video sample. This is
similar to the project that uses the webcam to measure heart rate. They created an
algorithm that amplifies the color to notice small changes, which occur on a heart beat. This
algorithm could also amplify small movements such as a pulse on someone’s wrist or the
breathing of an infant. This could be used in our project to amplify color and notice the pulse
of the subject .
It is also needed to be shown that heart rate variability relates to stress. A study was
done comparing the heart rate variability of people while at rest and while doing a mental
task that could be considered added stress. In this study, it is shown that with the added
stress, heart rate variability goes down which indicates stress .
Project Realistic Constraints
Wearability - The electronics need to be designed so that they can be worn
throughout the day. Depending on the electronics needed to take the
measurements, store this data, power the device, and send the data to the
smartphone, this may not be portable in the first prototype.
Data Storage - The amount of data that can be stored on the device might be a
constraint depending on how often the data is stored.
Battery - The amount of power that the device takes may not allow it to be worn
throughout the day because of the storage device and frequency that data is stored.
Battery should last a full day so that it can be charged at night.
Should accurately measure irregular heart rate on the device to accurately give a
Memory should store some data without being connected to smart phone in case
phone battery dies or is not in the immediate area.
Device should be wearable throughout the day to compete in the wearable
Should be sturdy enough to handle the outside environment and not need replaced
Communicate via Bluetooth to smartphone with application.
Application displays data and performs psychological questionnaire.
The system should be easy to use to attract the most users.
The device should cost less than $300 to compete with other devices on the market.
The device should not cause any harm to the user for long term use.
Should be attractive and stylish so that users will want to wear the device.
There are many factors that impacted the designs that are chosen. A tree of these
objectives is shown in Figure 1. The most important objective is that this device needed to
get some sort of Physiological data. It is decided that heart rate variability is much more
important than blood pressure and could be more accurately measured. Another very
important objective is the device wearability. If the device is not wearable, it is no different
than an ECG machine or Blood pressure monitor that is found in a Hospital or doctors
office. As technology increases, it is also important to be able to link to users smartphones
and display the health data to the user. The application should contain a Psychological
questionnaire, physiological data meaningfully, and be compatible on Android devices. The
last main objective is for the device to consume low power and have a battery that can last
throughout the day.
Figure 1: Objective Tree
To decide what project to implement, a decision matrix was created. This is shown in
Table 1. The first project that is compared is a Smart Shirt that performs a heart rate
variability reading. The second project is a Wrist worn Heart Rate and Blood Pressure
Monitor. The decision matrix shows that the Smart Shirt with a heart rate monitor is the best
option. The aim of this project is to create a product that can be worn throughout the day
without being bulky. We also want to be accurate in our measurements, which is where
many mobile blood pressure monitors and wrist heart rate monitors struggle.
Both of the devices are priced to cost about the same amount. Because the weight
of the smart shirt is distributed across your whole body and blood pressure hardware is very
heavy and awkward, the Smart shirt would have a large advantage in the weight category.
When it comes to accuracy and wearability, the smart shirt has a large advantage. It can be
easily worn all day and get accurate results, while a blood pressure monitor is awkward to
wear all day. Both of the devices are fairly even on power consumption, which is not one of
our main concerns.
Table 1: Alternative Solutions Decision Matrix
Cost Weight Accuracy Wearability Power
Smart Shirt with Heart Rate
Wrist Heart Rate and Blood
The Design Solution
The solution that is used is a wearable article of clothing that connects to a
weatherproof device. This device contains filtering circuitry and an Arduino Uno clipped to
the user’s belt or waistband. The Arduino is equipped with a Bluetooth module that
communicates with the user’s Android powered device. The device has the application that
is developed to communicate with the Arduino. Data is sent between the Arduino and the
Android smartphone, and the smartphone uses the data to create a physiological stress
rating. The app also offers psychological stress measurements in the form of a 10 question
survey. The user answers questions pertaining to their life, work, and overall fitness level,
and the survey gives them a rating. An equation is used to combine the psychological
score with the physiological score, and this score is displayed to the user. The user can
then use this to modify parts of their day in order to lessen their stress throughout the day.
This solution is chosen because of the expanding market for wearable health
devices. Smart clothing that takes physiological data is also becoming more popular. The
important component of the filtering circuitry is the ADS1293, which is a great analog front
end for mobile ECG applications. This signal processor filters out the noise generated
along with the heart signal and amplifies the signal from the heart. The Arduino Uno is
selected because it is well known microprocessor and is relatively easy to program and use
for any application, and the Bluetooth communication protocol is used because of its
simplicity and easy connection with smartphone devices.
The designed solution can be broken down into numerous different layers. The
highest level block diagram is shown in Figure 2. The entire system takes in electric signals
from the Smart Shirt and user input to the application. The system then outputs
physiological and psychological data on the application in the form of a combined stress
Figure 2: Block Diagram Level 0
Table 2: Stress Monitor Module Description
-Electronic Signal from Shirt
-App interaction from user
-Physiological information on Application
-Psychological information on Application
Functionality -Convert information from user and electronic shirt into statistics displayed
through application on android smartphone
In the second level of the system, the stress monitor is broken into two parts. This is
shown in Figure 3. The first part is a psychological survey which is implemented using a
mental survey. The other part of the system is an irregular heart rate monitor that takes in
the signal from the smart shirt and outputs a mental stress value.
Figure 3: Block Diagram Level 1
Table 3: Psychological Survey Module Description
-Answers to Questionnaire
-Level of emotional Stress
Functionality -Take answers to questions and convert them to a psychological stress
level inside of the android application that is created
Table 4: Irregular Heart Rate Monitor Module Description
Irregular Heart Rate Monitor
-Electronic Heart Signal from Shirt
-Level of Physical Stress
Functionality -Take signal from shirt, filter it, and send through Bluetooth to application
and display graphically.
The final layer of this design expands the details inside of the Irregular Heart Rate
Monitor and is shown in Figure 4. The first part consists of the smart shirt which gets data
from the electrodes. The smart shirt is then connected to the filtering circuitry unit (Texas
Instruments ADS1293). The circuitry unit then connects to the Arduino Uno that sends the
data over Bluetooth to the smart phone application. The data is then output to the user.
Figure 4: Block Diagram Level 2
Table 5: Sensoria Smart Shirt Module Description
Sensoria Smart Shirt
-Signal from electrodes on Sensoria Fitness Shirt
-Signal to nodes on front of shirt
Functionality -Convert signal from electrodes on shirt to the nodes on the front of the
shirt, ready to be transmitted.
Table 6: ADS1293 Module Description
-Signal from nodes on front of shirt
-Filtered Data to Arduino Uno
Functionality -Filter data from shirt and output to the Arduino
Table 7: Arduino Uno Module Description
-Filtered signal from ADS1293
-ECG Data to Bluetooth Module
Functionality -Take data from ADS1293, save it, and send it to the Bluetooth module.
Table 8: Bluetooth Device Module Description
-Data from Arduino Uno
-Signal to Smartphone over Bluetooth
Functionality -Take data and use Bluetooth classic to transmit data to smartphone
Once all of the data is sent between the units inside of the system, calculations need
to be done on this data. This design uses Heart Rate Variability and a Mental Survey to
Heart Rate Variability (HRV) measures the amount of time between beats and how
this changes in stressful and restful conditions. Figure 5 shows an example of HRV
intervals. Two different HRV calculations are used to calculate physical stress levels. The
first is called the mean RR value. This calculates the average amount of time between
heartbeats. Example RR values are also shown in Figure 5. The second calculation that is
done is called the pNN50 value. This value calculates the number of consecutive RR values
that differ by 50ms or more. For Example, if beat 1 occurs at time 0 seconds, beat 2 occurs
at time 0.5 seconds, and beat 3 occurs at time 1.06 seconds, there is a pNN50 occurrence.
This data leads to RR values of 0.5 and 0.56 which differ by more than 50ms.
Figure 5: Heart Rate Variability Diagram
There have been numerous studies that describe how stressful tasks affect heart
rate variability. One set of data is shown in Table 9 below. When performing a mental task,
the mean RR value decreases and your mean pNN50 percentage decreases.
Table 9: Heart Rate Variability Study Data
Mean RR (ms)
Mean pNN50 (%)
18.6 (± 14.8)
14.2 (± 12.6)
This application takes baseline readings while the user is at rest, and compares
these values to values that are obtained at stressful times. To calculate the Physiological
stress value, the equation in Figure 6 is used. The baseline values described above are
compared to the current values, scaled to give a value between 0 and 50, and divided by
the maximum acceptable Mean RR and pNN50 value. This equation gives a value between
0 and 100 when added together and is output in the app for the user to view.
Figure 6: Physiological Stress equation
The other part of the overall stress statistic is the mental stress data. The survey that
is taken consists of 10 questions and outputs a value between 0 and 100. This survey can
be taken once a day, but is not mandatory. If the user does not take a survey for numerous
days, the majority of the total stress value comes from physical data. Table 10 shows how
this data is scaled based on how long it has been since the user took the survey. After 6
days of not taking the survey, the mental data is not used at all.
Table 10: Stress Equations
Time Since Survey Was Taken (days)
Stress = PS/2 + MS/2
Stress = (3 PS)/4 + MS/4
Stress = PS
After all of the calculations are completed, the data needs to be output to the user.
Figure 7 shows a group of sample screenshots from the Android application. The first
shows the main menu of the app, which shows the Mental, Physiological, and Total stress
values. The next two screenshots show a sample survey question, and the survey results
screen which show your results both numerically and on a graph slide bar.
Figure 7: Android Application Screenshots (Main Page, Survey Question, Survey Results)
All of this information came together to assist in creating a design. All of these parts
are designed, but still need to be integrated, implemented, and troubleshooted.
All of the parts that have been designed need to be integrated together to produce a
working prototype. The first part of the system that connects straight to the smart shirt is the
ADS1293 filtering circuitry. The schematic of this integrated circuit is shown in Figure 8.
The ADS1293 is an analog front end for mobile ECG applications. It contains an
Electromagnetic Interference (EMI) filter at the input of each channel to filter out grid power
frequencies that normally operate around 60 Hz. A flexible routing switch is available to
switch between different channels.
The flexible routing switch outputs into the Instrumentation Amplifier for each
channel. This amplifier can increase the differential input voltage by ±400 mV. It can also
be switched between a low power mode or high resolution mode. The high resolution mode
has less noise than the low power mode at the cost of increased power consumption.
The next stage is the Sigma-Delta Modulator (SDM). This converts the output signal
from the Instrumentation Amplifier into a high resolution bit stream that can then be
processed by digital filters. The SDM can be configured to operate at 102.4 or 204.8
kHz. Operating at a higher frequency improves the resolution of the bit stream by
oversampling the signal at the cost of higher power consumption.
The last crucial stage of the ADS1293 signal processor is comprised of three
digital filters. The programmable digital filters reconstruct the signal from the SDM, and
each stage contains a fifth order SINC filter. Each stage further decimates and filters the bit
stream. The third SINC filter decimates the bit stream the most and is used to output an
accurate ECG signal. Every stage can be programmed to increase or decrease power
consumption and to alter the signal to noise ratio .
Figure 8: ADS1293 Functional Block Diagram 
Once the data is filtered by the ADS1293, it is sent over SPI (Serial Protocol
Interface) to the Arduino Uno. The timing diagram for this communication is shown in Figure
9. For SPI communication, there has to be a master (Arduino) and a slave (ADS1293). The
clock (SCLK) is output from the Arduino and input to the ADS1293 which synchronizes the
transaction. When the Arduino wants to receive data, it sets the Chip Select bit (CSB) LOW
and to finish the transaction, it sets the CSB HIGH. The transaction consists of 16 bits of
data. The first bit tells the ADS1293 whether we will be writing or reading. The next 7
specifies the read or write address. For our implementation, we are reading 3 bytes of data
from location 0x37, 0x38, and 0x39. The last 8 bits is the data that is read from the address
Figure 9: SPI protocol from ADS1293 (From TI)
Once the data is received on the Arduino, it is sent over Serial Bluetooth to the
smartphone app. On the Arduino side, the data is simply output using a Serial output
command. From there, the BlueFruit EZ-Link Bluetooth module does all the work. On the
Android application, code was written to read the device name and MAC address. Once
these values are received, the Bluetooth connection is made and data is constantly sent
over this connection.
Figure 10: Current Prototype
All of these modules are integrated together to form a complete prototype design.
This design is shown in Figure 10. This device is not currently portable, but with the
manufacturing technology to be able to place the ADS1293, microprocessor, and Bluetooth
module on an integrated circuit together, the device would decline in size dramatically.
The design that is integrated must be verified via testing and supporting figures. The
first test is to determine what kind of signal could be generated from the Sensoria smart
shirt connected to the ADS1293 Evaluation Module onto a personal computer via a USB
and supporting software. The smart shirt is used to collect the heart pulses from the human
subject and the leads are the inputs of the evaluation module. The software allows the
developer to program the ADS1293, allowing flexibility in design. After changing amplifier
gains, decimation rates, and modulator frequencies, a configuration is set up to attain the
ECG signal in Figure 11. The signal is over approximately a seven second interval and is
measured with respect to input voltage.
Figure 11: ECG Graph (Texas Instruments Software)
This is an important baseline test to see if an ECG signal could be attained using a
the Sensoria smart shirt with the ADS1293 analog front end signal processor. If this test did
not result in the figure above, the design cannot be successfully completed.
The next test is to connect the output of the ADS1293 to the Arduino Uno via Serial
Peripheral Interface (SPI). A digitized signal comparable to an ECG signal is desired to
output from the Arduino serial monitor. After receiving the data over SPI, Figure 12 is
obtained by graphing the Arduino Serial output in Excel.
Figure 12: ECG Graph (Excel data from Arduino)
The figure above is similar in form to Figure 11 from the TI software. The signal here
is digitized similarly to an ECG signal that is desired. This test verifies that data is being
sent via SPI protocol from the ADS1293 signal processor to the Arduino Uno
Finally, it must be verified that data being buffered in the Arduino Uno can be sent
via a Bluetooth serial module to a smart device. The same signal acquired in Figure 12 is
desired. The data received on the smart device must then be structured to represent an
ECG signal. This signal is shown below in Figure 13. This figure represents the data sent
from the Arduino Uno to the smart device.
Figure 13: ECG Graph (Android Application)
The last test that is done is to have the data sent to the smartphone, which would
use the physiological stress equations to calculate a value. The calculations are done on
the smartphone in the Android application on a separate thread. The application allows you
to save baseline data over a one minute period, and compare it to live data over a one
minute period. This gives a value for physiological stress. When this test is performed, it is
seen that the value is much more accurate when the user is sitting still, which could be a
problem to make this device wearable and accurate throughout the day.
The unit variable cost for the design is shown in Table 11. These parts were ordered
in November 2014 and these prices were at that time. Prices are likely to change as some
devices become more common.
Table 11: Device Unit Cost
Bluetooth Serial Link
Additional Electronic Components
Manufacturing Labor ($15 per hour)
The fixed costs for the design are shown below in Table 12. These are estimated
wages that are paid to developers to make this product ready to send onto the market.
These wages are split into 4 main categories shown in the table.
Table 12: Development Costs
Price (Million $)
Testing and Troubleshooting
A break-even analysis is used to find the approximate units that are needed to break
even after fixed costs and unit costs. For fixed costs, development of the facilities needed
to make the units approximate to around $10 million. For a better approximate unit cost,
$200.00 is used. If the sale price for each unit is $250.00, then the number of units needed
to break even is approximately 200,000 units.
Table 13: Work Breakdown Structure
be used, decide
what is needed,
Figure 14: Gantt Chart Graphical Representation
The wearable technology market is growing very quickly and new devices are
coming into the market very often. Smart clothing is one of the newest forms of this
technology to hit the market, with stress being an important function to monitor throughout
the day. The stress monitor that is developed uses a smart shirt to get heart beat readings,
which gives users a stress reading by calculating variable heart rate data. If this device is
continued and professionally manufactured, it could sell for a reasonable price and
accurately inform the users.
 Walker, Shane. "Wearable Technology - Market Assessment." IHS Electronics & Media.
September 2013. Web. 06 September 2014.
 "The Impact of Stress: 2012."http://www.apa.org. N.p., n.d. Web. 30 Sept. 2014.
 "Smart Clothes." healthinformatics -. N.p., n.d. Web. 1 Oct. 2014.
 Camm, John. "Heart Rate Variability." Circulation. American Heart Association, n.d.
Web. 1 Oct. 2014. <http://circ.ahajournals.org/content/93/5/1043.full>.
 "Measurement of heart rate variability: a clinical tool or a research toy?." JACC Journals.
Journal of the American College of Cardiology, 1 Dec. 1999. Web. 1 Oct. 2014.
 Corrales, Marina , Blanca Torres, Alberto Esquivel, Marco Salazar, and Jose Orellana.
"Normal values of heart rate variability at rest in a young, healthy and active Mexican
population ." SciRes 4.7 (2012): 377-385. Print.
 Poh, Ming-Zher, Daniel McDuff, and Rosalind Picard. "Non-contact, automated cardiac
pulse measurements using video imaging and blind source separation.." Division of Health
Sciences and Technology 18.10 (2010): 1-10. Print.
 "Technology: Innovation in‚ every interaction.." Apple. N.p., n.d. Web. 30 Sept. 2014.
 Wu, Hao-Yu, and Michael Rubinstein. "Eulerian Video Magnification." Eulerian Video
Magnification. MIT, 1 Feb. 2014. Web. 1 Oct. 2014.
 Taelman, J., and S. Vandeput. "Influence of Mental Stress on Heart Rate and Heart
Rate Variability - Springer." Influence of Mental Stress on Heart Rate and Heart Rate
Variability - Springer. Version 22. IFMBE Proceedings, 1 Jan. 2008. Web. 1 Oct. 2014.
[Snas602C, Texas Instruments Incorporated. ADS1293 Low Power, 3-Channel, 24-Bit
AFE for Biopotential Measurements (Rev. C) (2014): n. pag. Web.