Diagnostico

Made by srimanth / Artificial intelligence / Communication / Displays / Health / IoT

About the project

This is an intelligent patient monitoring solution that uses BLE to collect data from patients and a unique ML core that predicts body parameters like cholesterol using non-invasive techniques.

Project info

Difficulty: Easy

Platforms: Arduino, Cypress, Everything ESP, Microchip, Arm

Estimated time: 3 hours

License: MIT license (MIT)

Items used in this project

Hardware components

	STM32 Blue Pill	x 1
	36p 2.54 Mm Pin Male	x 1
	CC2541	x 1
	CR2302	x 1
	CR2302 holder	x 1
	The Ad8232 Sparkfun Single Lead Heart Rate Monitor Is A Cost-effective Board Used To Measure The Electrical Activity Of The Heart. This Electrical Activity Can Be Charted As An Ecg Or Electrocardiogram And Output As An Analog Reading. Ecgs Can Be Ext	x 1
	The Sparkfun Pulse Oximeter And Heart Rate Sensor Is An I2c Based Biometric Sensor Utilizing Two Chips From Maxim Integrated: The Max32664 Biometric Sensor Hub And The Max30101 Pulse Oximetry And Heart Rate Module. While The Latter Does All The Sens	x 1
	Passives	x 1
	Cy8ckit-062-ble Psoc 6 Pioneer Kit	x 1

View all

Software apps and online services

	Intel Dev Cloud
	Jupyter Notebook
	Anaconda
	Arduino IDE
	Processing IDE

Hand tools and fabrication machines

	Screwdriver	x 1
	Wrench	x 1
	Crimper	x 1
	Hot Glue	x 1
	Utility knife	x 1
	Soldering Station	x 1
	Flux	x 1
	Pen Flux	x 1
	Tweezers	x 1

View all

Story

Introduction

All around the world due to the COVID-19 pandemic many people are not getting the required medical attention. Doctors are not able to manage the medical resources and due to this, our current medical framework is collapsing.

In my place, Hyderabad, India many of the patients are not able to get admission into hospitals and they are in-home quarantine. But, if there was a way for a doctor to monitor these people remotely and with less effort, they would be able to help these people better.

Proposed idea

Using a simple wearable device and a software platform we plan to integrate the patient data into a platform and make processed data available for a doctor.

The product we designed collects patient data via a wearable, transmits them to a data aggregator via BLE, and then the aggregator pushes it to a server where a machine learning model generates inference. This along with the patient data is displayed on the doctor's computer.

Advantages

1) Simple and easy to use

2) Helps manage patients more efficiently

3) No extra efforts needed

Schematics, diagrams and documents

Low power BLE wearable with Spo2 , heart rate and ECG monitoring

CAD, enclosures and custom parts

Low power BLE wearable with Spo2 , heart rate and ECG monitoring

Go to download

Code

Max30102 STM32 Blue Pill interfacing code

Processing IDE interface

// Note : This snippet is not my own code 
// Credit : https://maker.pro/arduino/projects/how-to-build-ecg-heart-measuring-monitor-system

import processing.serial.*;

Serial myPort;        
int xPos = 1;         
float height_old = 0;
float height_new = 0;
float inByte = 0;

void setup () {
  
  size(1000, 400);
  println(Serial.list());
  myPort = new Serial(this, Serial.list()[0], 9600);
  myPort.bufferUntil('\n');
  background(0xff);
}
void draw () {
}

void serialEvent (Serial myPort) {
  String inString = myPort.readStringUntil('\n');
  if (inString != null) {
    inString = trim(inString);
    if (inString.equals("!")) {
      stroke(0, 0, 0xff); //Set stroke to blue ( R, G, B)
      inByte = 512;  // middle of the ADC range (Flat Line)
    }
 
    else {
      stroke(0xff, 0, 0); //Set stroke to red ( R, G, B)
      inByte = float(inString);
     }
  
     inByte = map(inByte, 0, 1023, 0, height);
     height_new = height - inByte;
     line(xPos - 1, height_old, xPos, height_new);
     height_old = height_new;

      if (xPos >= width) {
        xPos = 0;
        background(0xff);
      }
      else {
        
        xPos++;
      }
  }
}

Machine Learning Core

This notebook is hosted on intel dev cloud where we used the Intel(R) Xeon(R) Gold 6128 CPU @ 3.40GHz as the core compute instance. The model is basically a Decision Tree Regressor that was trained on data partially collected from a kaggle dataset and partially from real-life sensor data.

ECG Sensor interfacing

void setup() {
Serial.begin(9600);
pinMode(10, INPUT); 
pinMode(11, INPUT); 
}
void loop() {
if((digitalRead(10) == 1)||(digitalRead(11) == 1))
{
Serial.println('!');
}
else{
Serial.println(analogRead(A0));
}
delay(1);
}

Shop

Project Hub

Video

Blog

Diagnostico

About the project

Project info