This document discusses using machine learning and IoT to detect stress levels through heartbeat monitoring. It presents a system that uses a Node MCU development board and pulse sensor to collect heartbeat readings from a person. The readings are sent to a server for analysis and to build a fingerprint of the user's stress levels over time. The system can predict if a person is stressed or relaxed based on their heart rate. Testing showed an accuracy of 66-68% for this stress detection.
1. Machine Learning and IOT
for Prediction and Detection
of Stress
Guided By:
Prof. Ms. P. V. Deshmukh
Presented By:
Mr. Sumitkumar S. Sahare(75)
2. Contents
1. Introduction
2. Block diagram
3. Data flow diagram
4. Hardware component
5. Result and analysis
6. Comparing Result
7. Conclusion
8. References
3. Introduction
Remote stress detector is an IOT device which can detect the
stress level of person using his/her heartbeat reading.
This device locally collect heart beat reading from a person and
sends it to a server on Digital Ocean.
Node MCU is used as the development board and micro-python
for programming language.
we can identify a lot of things, for example whether the person
is nervous or not, whether the person is in apprehension or
fear, where the person is in apprehension or fear.
4. Internet of Things ?
IOT is a combination of hardware and software technologies
along with embedded devices which enables to provide
services and facilities to any one, anything, anywhere
required using any network.
The system can be accessed through a web portal by the web
administrator, physician, patients, medical staff and also
researchers.
5. Fig. The general architecture of the health care information system
Health care information system
8. Hardware Components
Fig. Node MCU
It establishes a connection with Wi-Fi and supports micro
python which is minimal version python.
It contain a Wi-Fi module
ESP8266 along with a 32
bit microprocessor
(Tensilica Xtensa LX106)
along with 160kb RAM.
Node MCU
9. It detects the pulse rate of the body which can counted per
unit to find out the heart rate.
It records a reading by
placing the tip of a finger
on it or attaching the
sensor with wrist.
Fig. Pulse Sensor
Pulse Sensor
10. Server and program flow
The node MCU periodically takes a heartbeat reading and
sends it, along with timestamp to our server.
The server has been implemented in flask and deployed on
a digital ocean droplet running Ubuntu 16.04 LTS.
The server assembles fingerprinted record of the user’s
heartbeat at different times of the day.
11. Result and Analysis
Stress labels are calculated and median in our data set was
stressed and map that to our detector.
The exercise
detection part of
our paper does
work well because
that relationship is
clearly defined as
50-80% of the
max heart rate.
12. fig. the person is exercising fig. the person is at rest or relaxed.
Determining whether the person exercising or rest
15. Conclusion
Based on heart beat we can predict whether a person is in
stress or not.
It can be used in our daily lives to monitor our anxiety and
stress levels which might help in better living.
16. References
[1] Purnendu Shekhar Panday “Machine Learning and IOT for Prediction and
Detection of Stress” 2017 17th International Conference on Computational Science
and Its Applications (ICCSA) DOI:10.1109/ICCSA.2017.8000018
[2] Iuliana Chiuchisan, Doru-Gabriel Balan, Oana Geman, Iulian Chiuchisan, Ionel
Gordin “A security approach for health care information systems” 2017 E-Health and
Bioengineering Conference (EHB)DOI: 10.1109/EHB.2017.7995525
[3] Cornelia Setz, Bert Arnrich, Johannes Schumm, Roberto La Marca, and Gerhard
Troster, Member, IEEE, and Ulrike Ehlert, “Discriminating Stress From Cognitive Load
Using a Wearable EDA Device”, IEEE transactions on Information Technology in
Biomedicine, Vol. 14, No. 2, March 2010.
[4] Sreekanth K U, Nitha K P “A Study on Health Care in Internet of Things”
International Journal on Recent and Innovation Trends in Computing and
Communication Vidya Academy of Science & Technology (VAST) ISSN: 2321-8169
Volume: 4 Issue: 2 044 - 047.