1 of 29

UG PROJECT PRESENTATION

Swasthshala-Cyber Physical System for Intelligent Health Care

Submitted by:

Astha Yadav(19095021) Guguloth Yakesh(19095038) Shourya Gupta(19095089) Yash Garg(19095123)

Under the Guidance of

Dr. N. S. Rajput

(Department of Electronics Engineering)

2 of 29

Executive Summary

Current situation of healthcare when compared with the top economies is very poor and number of doctors and hospital beds are very less than the recommended level.

Seeing at the major causes of death among children and adolescent, we can deduce that many deaths could have been prevented if their health were monitored regularly.

Government has launched some schemes to combat this problem and it has worked up to some extent in saving life and detection of fatal diseases in initial level in school students.

Swasthshala is our innovative way to bridge the gap between school students and healthcare.

Swasthshala is basically an application which can be used to maintain the data of regular health check-up of students of community school with facial recognition feature.

3 of 29

Current situation of Healthcare in India

No of physicians per 1000 people in India is below the standard set by WHO.
India spends a very low percentage of its GDP in healthcare when compared with other major economies.
As the second most populous country in the world with over 1.3 billion people, India also has to focus on building quality healthcare infrastructure and economically develop itself in such a way that its overall infrastructure is preventive against health risks.
Out of 189 countries, India ranks 131 in Human Development Index 2020 prepared by the United Nations Development Programme (UNDP).

Source: World Bank

4 of 29

Hospital beds in India are far below the recommended level and its spread is very uneven across the country.

India has an abysmally low 0.5 public hospital beds per 1,000 population and mere 1.4 beds, including public and private hospital beds per 1,000 persons.
The healthcare services in India face a shortfall of 3.5 million hospital beds in terms of public hospitals. The shortfall is reduced to 2.4 million beds with the inclusion of private hospitals.
The southern states of Karnataka, Telangana and Kerala have the maximum number of beds per 1,000 population (of public and private hospitals). On the other hand, states such as Bihar, Odisha, Chhattisgarh and Jharkhand lag behind and rank the lowest in hospital beds.
Not only does India sit at the bottom of the chart on healthcare spending among top 10 major economies covered in the analysis, it also has the lowest number of beds available at its public hospitals.
As many as 69 percent of hospital beds in India are concentrated in urban areas

Source: World Bank, Moneycontrol

5 of 29

If we look at major causes of death among school-age children and adolescents, we can see that many life lost could have been prevented, if at school level, there were some programme to regularly check-up on student’s physical and mental health.

Source:WHO

6 of 29

To combat the rising health issues among children and adolescent, the government has launched some schemes.

�Rashtriya Kishor Swasthya Karyakram (RKSK)

Objectives

Target Group

The key objective of this scheme is adolescent participation and leadership, Equity and inclusion, Gender Equity and strategic partnerships with other sectors and stakeholders.
The strategy envisions that all adolescents in India are able to realize their full potential by making informed and responsible decisions related to their health and well-being, and by accessing the services and support they need to do so.

The scheme focuses on age groups 10-14 years and 15-19 years with universal coverage, i.e. males and females; urban and rural; in school and out of school; married and unmarried; and vulnerable and under-served.

�

Strategies

Strengthening of Adolescent Friendly Health Clinics (AFHC)
Peer Education (PE)
Weekly Iron and Folic Acid Supplementation Programme (WIFS)
Menstrual Hygiene Scheme (MHS)

Source: Ministry of Home Affairs

7 of 29

Objective

Target Group

It is an initiative aiming at early identification and early intervention for children from birth to 18 years to cover 4 ‘D’s viz. Defects at birth, Deficiencies, Diseases, Development delays including disability.

The services aim to cover children of 0 -6 years of age in rural areas and urban slums in addition to children enrolled in classes I to XII in Government and Government aided Schools.
It is expected that these services will reach to about 27 crores children in a phased manner.

Health Conditions to be screened

Rashtriya Bal Swasthya Karyakram (RBSK)

Source: Ministry of Home Affairs

8 of 29

These government schemes have shown that with increasing the focus on health of children and adolescents through regular check-up and screening, we can improve and even save the life of many.

Source: Deccan Herald, Indian express, the tribune

9 of 29

Swasthshala is our innovative way to bridge the gap between school students and healthcare.

Swasthshala is basically an application which can be used to maintain the data of regular health check-up of students of community school.

About

Features

It is an application built with flutter framework which means it can be run on android, iOS and windows.
There is an option of login via facial recognition which is going to make the whole data collection process very smooth.
Django framework has been used for the backend of the app which makes this app very secure. This is very much needed as health data are very sensitive piece of information.
This app is highly scalable. Depending on the requirements, it can be scaled up or down.
ML algorithms have also been used to compare the data and detect anomaly.

10 of 29

The health data of the student is collected which is then fed into the student profile.

Student’s identity is recognized through facial recognition.

Data is then sent to a backend system. This will be stored on a cloud-based database for high availability.

Data once collected can be used for time series analysis on individual students and can be used to perform complex analysis with ML techniques.

Data and findings of the analysis of every student is reviewed by a doctor and if any anomaly is found, and doctor shall recommend the further steps.

The working of the Swasthshala looks like the following:

11 of 29

The Swasthshala App

This app will enable the periodic collection of student health data.
The collection of data will work as follows.

Each authorised person can upload measured parameters on the app.
The authorised person will click image of student. Images of all the registers students will be stored on the server. And any new student can be added easily.
The health data along with the photograph will be sent to the backend.
The backend will perform a matching algorithm to identify the person in image and add those records in table with the identified person.
The admin can view/download the real time updated data anytime from the admin site of backend.

12 of 29

Add new student

Enter record for registered students

User Experience in the app

13 of 29

1. Arduino Uno R3

Arduino Uno R3 is a kind of ATmega328P based microcontroller board.
It has 14 digital I/O pins. From these 14 pins, 6-pins can be utilized like a PWM output.
This board includes 6 Analog inputs, a USB connection, quartz crystal of 16 MHz, a power jack, an ICSP header and an RST button.
The memory includes 32 KB of FLASH and 0.5 KB memory is used for the Boot loader.

Design Details

14 of 29

2. ESP8266 Wi-Fi module

ESP-8266-01 is a Wi-Fi module which enables the Micro-controller to easily access the Wi-Fi network. To add network connectivity to the arduino board.
This module requires bare minimum internal circuitry.
It’s entire solution, including the front end module is designed to occupy as low as possible PCB area as well.

15 of 29

3. Max-30102 Sensor

To measure heart beats and SPO2.
It includes internal, photodetectors, LEDs, low noise electronics and optical elements with ambient light rejection feature.
It uses a technique called photoplethysmography (PPG).
Tiny 5.6mm x 3.3mm x 1.55mm optical module with 14-Pins.
Ultra-Low Power Operation which aids in handling Mobile Devices.

16 of 29

4. GSR Sensor

Galvanic Skin Response
Used to measure electrical conductance of the skin
Consists of two electrodes which can be put on the two fingers of one hand
Change in conductivity of skin denotes the change in state of emotions

17 of 29

5. LM-35 Temperature Sensor

A precision Integrated Circuit(IC) temperature sensor with its output proportional to the temperature.
Temperature is measured more accurately than it could be done with any thermistor.
The operating temperature range of this sensor is from -55°C to +150°C.
It has low output impedance, a linear output, and precise inherent calibration.

18 of 29

6. Weight (Load Cell)

The load cell is used for measuring the forces applied. Can be used a weight sensor.
The load cell combined with a HX711 IC can be used a weight sensor.

19 of 29

7. LCD Display

It is a specific type of electronic display module that is used in several places.
It has a LCD with 162 characters and a parallel interface.
It has the capacity to read from and write to the display.
The main advantages of using this module are it’s low cost and ease of programming.

20 of 29

8. BP measurement

Most important vital sign is blood pressure. But measuring it is complicated and requires lot of calibration and needs inflating equipment with an accurate timing sensors for heart beats.
There exists a novel method that can estimate Blood pressure using the output from the photodetector of PPG.
This uses PWTT. It is the time interval between two pulses measured on the same artery. More specifically, PWTT is the time interval between the R peak of the ECG and the minimum peak of the PPG.

21 of 29

Measurement of PWTT

22 of 29

BP measurement

A paper [4] was found where blood pressure analysis was performed and they mentioned a linear regression that estimated the blood pressure using PWTT

The systolic blood pressure is given by

SBP = −0.6881× PWTT + 228.591

23 of 29

Block Diagram

ARDUINO

Temperature sensor

GSR Sensor

Heart Rate Sensor and SPO2

ESP8266 Module

BACKEND (cloud)

I2C Interface

Analog Pin Interface

Weight Sensor

Digital Pin Interface

24 of 29

Data analysis and visualization

The recorded health data can be exported to a form of CSV.
Also the Django framework we used provides an easy to use admin panel to display the data in tabular form.
The admin panel can also be used for user and student management.

25 of 29

Time Series analysis of data of patient/student data

The data collected from the application will be used to perform time series analysis of student’s health.

Various benefits of such analysis will be

Predict outbreak of any disease in a particular region.
Ensure proper nutrition and prevent malnourishment of young children.
Early diagnosis and treatment of severe diseases and disorders that are easily curable at early stages.

The data and its insights will be shared and discussed in close observation with doctors to ensure necessary steps are taken.

26 of 29

We have tried to perform analysis using data publicly available on the internet.

Many problems like obesity, underweight and overweight problems can be directly observed from height, weight and BMI metrics.

Other problems that are not very easily visible unless observable symptoms start appearing require much in-depth analysis of available data. This can be performed by use of state-of-the-art Machine Learning and Deep Learning methods.

We can user LSTMs for time series analysis, ANN and DNN techniques combined with time-based analysis to reveal complex patterns that are not visible to human eyes.

Time Series analysis of data of patient/student data (Continued)

27 of 29

Graphical Representation of the data accumulated

28 of 29

Future Scope

Implement on the institute level.
Add more health indicators and sensors. To gain more insight into health data.
Have a more compact and portable design.
Can be made into a Smart Kiosk which will automatically record all the health data of a person and send it to cloud.
Train ML models on collected data to identify early signs of serious diseases.