DrCloud is a system built with an Intel Galileo board that records heart rate, temperature, other vitals, and GPS location using Arduino sensors and places a marker on an intelligent map using the Google Maps API.
This allows doctors the ability to see critical geographical areas in terms of the area's health and average disease rate.
This works! we have structure tested for multiple Galileo boards and creating pindrops.
The GPS is accurate and the Galileo is powered easily by 5 volts of power. Even a hand-crank generator could run the system.
Since both the Intel Galileo and the sensors are inexpensive, this is feasibly scalable. With more galileo boards, the data is better and better. Same with any sensors. Internet connection would need to be part of a scalability solution as well. This could be solved using local internet connections, or in other cases satellite-based internet setups. In addition, data could be stored locally and collected at a central internet-connected machine to still produce a map in a poorly-connected area.
We simply think that DrCloud has the potential to be an invaluable tool for the future of triage in disaster and disease occurrences.
Thanks,
James Carr, Paul Moulton, James Goff, David Chen, Don-Wook Shin
DrCloud Team
Here are some details on the implementation:
http://github.com/pmoulton/galileo-hack
main.py- uses python mechanize library (required separate installation on our galileo web server)
Input: our IoT cloud account
output: DoctorCloud.xls
auto-fills the https form on the IoT website and clicks "submit"
takes a url computed by drcloudscript.sh (below) and downloads an XLS in the authenticated IoT web session
pypypy.py - uses xlrd and another json library in python to convert the downloaded xls from IoT into json
input: DoctorCloud.xls made by main.py
output: data.json made after converting xls to json
script.py - parses the json from pypypy.py and turns it in to a single, manipulated python list that the javascript map webpage can access
input: data.json from previous script ^
output: info.js with filterred data
we average all of the data points from the last 7 minutes so that it is the most reliable. we ran a logarithmic algorithm on the temperature sensor, i believe (or we were considering it to get cleaner data)
info.js has four numbers: lat, lon, temp, and heart rate
these are real geographical numbers that are used on the map
index.html - javascript webpage that accesses the coordinates and sensor values for a specific point.
input: js/info.js
output: html webpage online
drcloudconverter.sh- ties all of the above together
once python mechanize in main.py has logged in the session, main.py checks stdin from this script for a url. here is how we build the url:
the url for the XLS is somewhat predictable. We wanted all four of our IoT data variables, from the last 5 seconds, in an xls.
once authenticated, we passed the 4 variables into the url and the current UTC time as the end date, with 400 seconds (about 7 minutes) previous as the start date. This gives us 7-minute relevant data, overwritten to DoctorCloud.xls, accessed on page refresh.
Runs the above commands in succession, they all work together, calling each other's output files
runs in the /var/www directory of our galileo web server and makes the final files available to CloudDr.html, which displays the final map.
in the sample_files folder are some of our most recently-computed versions of the passthrough files. This is so you can see what they look like, etc.
The scripts above are all in scripts folder in the github.