Patient Monitoring System using 96Boards

This repository consists of instructions to set up a Patient Monitoring Systems using 96Boards with the help of BLE Mesh network.

Table of Contents

• 1) Hardware
  • 1.1) Hardware Requirements
• 2) Software Setup
  • 2.1) Setting up the Mesh Nodes
  • 2.2) Setting up the Cloud Service
  • 2.3) Setting up the Gateway
• 3) Hardware Setup
• 4) Patient Monitoring System
• 5) Video Demonstration

1) Hardware

1.1) Hardware Requirements

• Dragonboard410c
• 96Boards Compliant Power Supply
• 96Bords Carbon
• TMP112 Sensor
• CCS811 Sensor
• Push Button
• Connecting wires

2) Software Setup

Below steps are used to setup the nodes and Gateway in BLE Mesh network. In this project 2 Carbon boards will be used as Server nodes and one Carbon board will be used as a Client node. Dragonboard410c will be used as a Gateway.

2.1) Setting up the Mesh Nodes

Following steps are used to setup the BLE Mesh nodes:

Programming nRF Co-Processor

[Execute this section on all Carbon boards using Host machine]

Follow the below steps to flash the hci_spi sample application to the nRF co-processor on 96Boards Carbon using a Linux enabled host machine.

First setup the development environment as mentioned here.

$ git clone https://github.com/Mani-Sadhasivam/zephyr.git
$ cd zephyr
$ git checkout ble_mesh_3
$ source zephyr-env.sh
$ samples/bluetooth/hci_spi
$ mkdir build
$ cd build
$ cmake -DBOARD=96b_carbon_nrf51 ..
$ make

Once the application is successfully built, flash the binary onto the nRF chip by following this guide. Now, this chip can provide HCI interface to STM32 via SPI.

Programming Sensor Server application

[Execute this section on the Carbon boards to be used as Server nodes using Host machine]

Next step is to program the Sensor Server application to the STM32 chip on the Carbon boards to act as a Sensor Servers.

Move to the top of the cloned Zephyr repository.

$ cd zephyr
$ cd samples/bluetooth/ble_mesh_srv
$ mkdir build
$ cd build
$ cmake -DBOARD=96b_carbon ..
$ make

Note: You should set different NODE_ID in ble_mesh_srv/src/main.c for different Server nodes. For this project, it is assumed that first node contains NODE_ID of 1 and second node contains NODE_ID of 2.

Now, the built binary can be flashed by following this guide. Once the application binary is flashed, connect the UART port of Carbon board to the Host machine using USB-A to USB-B Micro cable and bring up the serial emulation tool like minicom on the corresponding port.

After board reset you should see the below message on the serial port:

Initializing...
Bluetooth initialized
Mesh initialized

Programming Sensor Client application

[Execute this section on the Carbon board to be used as a Client node using Host machine]

As like the Sensor Server application, Sensor Client application is also need to be programmed on another Carbon board.

Move to the top of the cloned Zephyr repository.

$ cd zephyr
$ cd samples/bluetooth/ble_mesh_cli
$ mkdir build
$ cd build
$ cmake -DBOARD=96b_carbon ..
$ make

Now, the built binary can be flashed by following this guide. Once the application binary is flashed, connect the UART port of Carbon board to the Host machine using USB-A to USB-B Micro cable and bring up the serial emulation tool like minicom on the corresponding port.

After board reset you should see the below message on the serial port:

Initializing...
Bluetooth initialized
Mesh initialized

Provisioning and Configuring the Nodes

[Execute this section on all Carbon boards using Host machine]

Next step is to provision and configure both the nodes to be used in BLE Mesh network. This requires the meshctl utility in Bluez stack. Install Bluez by following the below steps.

$ git clone https://git.kernel.org/pub/scm/bluetooth/bluez.git
$ cd bluez
$ bootstrap
$ ./configure --prefix=/usr --mandir=/usr/share/man --sysconfdir=/etc --localstatedir=/var --enable-mesh
$ make
$ sudo make install

Above commands will install Bluez on the host machine with BLE mesh support.

Server Node

[Execute the below section on the Carbon boards to be used as Server using Host machine]

Now, execute the below steps for provisioning and configuring the Server node:

$ cd bluez/mesh
$ meshctl

This will bring up the meshctl command prompt.

[meshctl]# discover-unprovisioned on
[meshctl]# provision dddd

Here we are provisioning the node dddd which is default address for all un-provisioned nodes.

Now, enter the OOB number displayed in the Server's serial terminal here.

[meshctl]# menu config
[meshctl]# target 0100

Here 0100 is the node address after provision. Modify it based on the assigned address.

Now, generate AppKey and bind it to the 0th element of models 1100 and 1000.

[meshctl]# appkey-add 1
[meshctl]# bind 0 1 1100
[meshctl]# bind 0 1 1000

Next, add the subscription and publish info to the models.

[meshctl]# sub-add 017a c000 1100 
[meshctl]# pub-set 017a c000 1 0 0 1100 
[meshctl]# pub-set 017a c000 1 0 0 1000 

Once, all the above commands are executed successfully, we can assume that the Sensor node is configured.

Client Node

[Execute the below section for the Carbon board to be used as Client using Host machine]

Now, execute the below steps for provisioning and configuring the Client node:

$ cd bluez/mesh
$ meshctl
[meshctl]# discover-unprovisioned on
[meshctl]# provision dddd

Now, enter the OOB number displayed in the Client's serial terminal here.

[meshctl]# menu config
[meshctl]# target 0101

Here 0101 is the node address after provision. Modify it based on the assigned address.

Now, generate AppKey and bind it to the 0th element of Client Sensor model (1102).

[meshctl]# appkey-add 1
[meshctl]# bind 0 1 1102
[meshctl]# bind 0 1 1001

Next, add the subscription and publish info to the models.

[meshctl]# pub-set 017b c000 1 0 0 1102
[meshctl]# sub-add 017b c000 1102
[meshctl]# sub-add 017b c000 1001

Once, all the above commands are executed successfully, we can assume that the Client node is configured.

2.2) Setting up the Cloud Service

We are going to use ThingSpeak as the cloud service provider and Twilio for sending emergency alerts.

Setting up Thingspeak

Using Thingspeak is very straightforward and easy task. Just create an account on their website then sign-in and follow the below steps:

1. After Signing in, you will be provided with a window having a button named New Channel. Click it and it will proceed to create a new channel in Thingspeak.

2. In the next window, provide the following information to create our channel:

• Name: Patient 1
• Description: Dashboard for Patient 1
• Field 1: Temperature
• Field 2: Co2
• Field 3: Emergency Button

Then click, Save Channel to save the channel information.

Note: Note down Write API Key, Read API Key and Channel ID for future use.

3. Now our channel is created and we have declared 3 fields: Temperature, Co2 and Emergency Button. These are the parameters which will be streamed from the BLE mesh network.

4. Next step is to add visualizations for the created fields. Goto Apps->Plugins. Click New. There will be 3 options for creating the visualization, for this project we will be using Google Gauge. So, select it and click Create.

5. In the next window, enter Name as Temperature and replace the Javascript code with below so that we can create a Google gauge for Temperature readings:

<script type='text/javascript' src='https://ajax.googleapis.com/ajax/libs/jquery/1.9.1/jquery.min.js'></script>
<script type='text/javascript' src='https://www.google.com/jsapi'></script>
<script type='text/javascript'>

  // set your channel id here
  var channel_id = 404689;
  // set your channel's read api key here if necessary
  var api_key = 'A173BFJ51PIY91RZ';
  // name of the gauge
  var gauge_name = 'Temp';

  // global variables
  var chart, charts, data;

  // load the google gauge visualization
  google.load('visualization', '1', {packages:['gauge']});
  google.setOnLoadCallback(initChart);

  // display the data
  function displayData(point) {
    data.setValue(0, 0, gauge_name);
    data.setValue(0, 1, point);
    chart.draw(data, options);
  }

  // load the data
  function loadData() {
    // variable for the data point
    var p;

    // get the data from thingspeak
    $.getJSON('https://api.thingspeak.com/channels/' + channel_id + '/feed/last.json?api_key=' + api_key, function(data) {

      // get the data point
      p = data.field1;

      // if there is a data point display it
      if (p) {
        displayData(p);
      }

    });
  }

  // initialize the chart
  function initChart() {

    data = new google.visualization.DataTable();
    data.addColumn('string', 'Label');
    data.addColumn('number', 'Value');
    data.addRows(1);

    chart = new google.visualization.Gauge(document.getElementById('gauge_div'));
    options = {width: 220, height: 220, max: 80, redFrom: 60, redTo: 80, yellowFrom:40, yellowTo: 60, greenFrom: 20, greenTo:40, minorTicks: 10};

    loadData();

    // load new data every 15 seconds
    setInterval('loadData()', 15000);
  }

</script>

Then, click Save

Note: Replace channel_id and api_key with your ID and key.

6. Next, repeat Step 4, 5 to create another gauge for Co2 readings.

7. Enter Name as Co2 Concentration and paste the below contents in Javascipt code editor.

<script type='text/javascript' src='https://ajax.googleapis.com/ajax/libs/jquery/1.9.1/jquery.min.js'></script>
<script type='text/javascript' src='https://www.google.com/jsapi'></script>
<script type='text/javascript'>

  // set your channel id here
  var channel_id = 404689;
  // set your channel's read api key here if necessary
  var api_key = 'A173BFJ51PIY91RZ';
  // name of the gauge
  var gauge_name = 'Co2';

  // global variables
  var chart, charts, data;

  // load the google gauge visualization
  google.load('visualization', '1', {packages:['gauge']});
  google.setOnLoadCallback(initChart);

  // display the data
  function displayData(point) {
    data.setValue(0, 0, gauge_name);
    data.setValue(0, 1, point);
    chart.draw(data, options);
  }

  // load the data
  function loadData() {
    // variable for the data point
    var p;

    // get the data from thingspeak
    $.getJSON('https://api.thingspeak.com/channels/' + channel_id + '/feed/last.json?api_key=' + api_key, function(data) {

      // get the data point
      p = data.field2;

      // if there is a data point display it
      if (p) {
        displayData(p);
      }

    });
  }

  // initialize the chart
  function initChart() {

    data = new google.visualization.DataTable();
    data.addColumn('string', 'Label');
    data.addColumn('number', 'Value');
    data.addRows(1);

    chart = new google.visualization.Gauge(document.getElementById('gauge_div'));
    options = {width: 220, height: 220, max: 2000, greenFrom: 500, greenTo:1200, redFrom: 1600, redTo: 2000, yellowFrom:1200, yellowTo: 1750, minorTicks: 10};

    loadData();

    // load new data every 15 seconds
    setInterval('loadData()', 15000);
  }

</script>

Then, click Save.

Note: Replace channel_id and api_key with your ID and key.

8. Now, we have created two Google gauges for displaying the Temperature and Co2 readings. Next step is to add those to our channel. Goto Channels->My Channels and select the channel Patient 1. In the channel dashboard, click Add Visualizations.

9. Select the following widgets:

• Temperature
• Co2
• Field 1 Chart
• Field 2 Chart

Finally, click Save.

Once all of the above mentioned steps are executed successfully we will have 4 widgets appearing on our channel.

Setting up Twilio

For sending the Emergency Alert, we will be using Twilio. Create an account there and follow the below steps:

1. Sign-in to the Twilio Console and note down the displayed ACCOUNT SID and AUTH TOKEN fields.

2. Now, verify a mobile number for which Twilio can send Emergency alerts.

3. Go to the Patient 1 dashboard we created on Thingspeak and select Apps->ThingHTTP->New ThingHTTP.

4. Enter the following credentials:

• Name: Patient Monitoring Alert
• URL: https://api.twilio.com/2010-04-01/Accounts//Messages.json
• HTTP Auth Username: Paste your Twilio ACCOUNT SID
• HTTP Auth Password: Paste your Twilio AUTH TOKEN
• Method: POST
• Content Type: application/x-www-form-urlencoded
• Body: From=+12x4xx5xxx7&To=+16x3xx4xxx8 &Body=Emergency Alert

Note: Enter your Twilio number in From= field and destination number in To= field.

Click Save ThingHTTP.

5. Next, click Apps->React->New React and enter the below information:

• React Name: Emergency Trigger
• Condition Type: Numeric
• Test Frequency: On Data Insertion
• Condition: Patient 1 3(Emergency Button) is equal to 1
• Action: ThingHTTP Patient Monitoring Alert
• Options: Run action each time condition is met

Click Save React.

Now, we are done with all settings needed to visualize the data from mesh network on cloud and also an Emergency alert when the push button is pressed

Note: You need to setup Thingspeak channel for all Server nodes used. Just replace the Read API key, channel ID and channel name accordingly. For this project, we need to setup two channels for two Server nodes.

2.3) Setting up the Gateway

Once the nodes are configured as mentioned above, the gateway needs to be setup in order to receive data from the client node via USB to UART. For this, we are going to use one python script which captures and uploads the data to cloud.

The following instructions assume that the Dragonboard410c(Gateway) has been setup properly with latest Debian release from Linaro.

First, install the required packages:

$ pip install --user pyserial
$ pip install --user thingspeak

Next, save the below contents to a file named pyserial.py

import serial
import sys
from time import sleep
import urllib2

key1 = 'Enter your Write key for channel 1'
baseURL1 = 'https://api.thingspeak.com/update?api_key=%s' % key1
key2 = 'Enter your Write key for channel 2'
baseURL2 = 'https://api.thingspeak.com/update?api_key=%s' % key2

# open serial
ser = serial.Serial('/dev/ttyUSB0', 115200)

while True:
    try:
        # Read serial
        ser_data = ser.readline()
        sen_data = ser_data.split(',')

	if (sen_data[0] == '1'):
            # Upload data to thinkspeak channel 1
            thing_status = urllib2.urlopen(baseURL1 + "&field1=%s&field2=%s&field3=%d" % (sen_data[1], sen_data[2], int(sen_data[3])))
            print thing_status.read()
            thing_status.close()

        if (sen_data[0] == '2'):
            # Upload data to thinkspeak channel 2
            thing_status = urllib2.urlopen(baseURL2 + "&field1=%s&field2=%s&field3=%d" % (sen_data[1], sen_data[2], int(sen_data[3])))
            print thing_status.read()
            thing_status.close()

        # Thingspeak only accepts data once every 15 seconds
        sleep(15)

    except KeyboardInterrupt:
        print "Exiting..."
        sys.exit()

Note: Enter your Write API key in key1/key2 variables with single quotes ('').

3) Hardware Setup

• Make sure the Dragonboard410c and Carbon boards are powered off
• Connect TMP112, CCS811 to I2C_1 and Push button to PC8 of Carbon board to be used as a server node.
• Repeat Step2 on all server nodes.
• Connect the Carbon board to be used as Client to Dragonboard410c through USB available at the UART port.

Note: At this point both the Carbon boards should be provisioned and the Gateway should be setup.

4) Patient Monitoring System

Once everything is setup as mentioned above, execute the python script on the gateway to deploy the patient monitoring system.

Now, the Dashboard on Thingspeak will get updated as below and if the emergency button is pressed, the registered user will get an emergency alert SMS.

5) Video Demonstration