Hey guys!
Dishant Shah here. This is the first blog post related to my Buzzbeat v2.0 project which also happens to be my official entry to this year's Design for a Cause 2021 challenge. I am grateful to the element14 community for choosing me as an official challenger and while I have not yet received my official kit, I have decided to start blogging starting right from the previous work and initiation of this concept/project which won't need the hardware.
Buzzbeat v1.0 is documented on this hackster.io project here. The main premise behind developing BuzzBeat was the following -
Doctors/nurses/staff has to continuously check the multi-para-monitor attached to the patient.
Multi-parameter monitors are costly and are only found in ICUs, especially in developing countries. Underdeveloped countries might not even have them in most hospitals
Normal cheap monitors or medical sensors measure only 1-2 parameters. Hence, people have to buy a lot of different devices just to measure common parameters.
Using these collected sensor signals to diagnose something using AI/DL.
So, we developed the BuzzBeat v1.0 prototype keeping the following features in mind -
- It will have an app and some notification system that can notify the doctors or nurses anytime, anywhere without them having to check the patient again and again.
- The notification should not be distractive and each notification should have some priority levels based on which the medium of notification would be selected. Eg. for high priority notifications vibrations could be used, for low priority ones, normal text messages could be sent, for medium-level ones emails could be sent for the doctor to have a record.
- The device will monitor multiple parameters together and harbor multiple sensors.
- The device will use some AI-based models to analyze those signals and parameter values and perform some level of diagnostics.
The parameters we are targeting are:
Body Temperature (BT)
SPO2 levels
Heart Rate (HR)
Heart Rate Variability (HRV)
Electrodermal Activity (EDA)
ECG waveform
PPG waveform
The parameters with a have been implemented in the BuzzBeat v1.0 and those with need to be added in BuzzBeat 2.0. Some more parameters can be measured like respiratory rate, blood pressure, tidal volume, etc. but that will make our system bulky and expensive. These parameters can be included later in a more sophisticated solution that encompasses all extra parameters.
The symptoms/diseases that we can identify potentially are -
Hypoxemia (low SPO2 levels)
Arrhythmia (Irregular Heart Beats)
Atrial Fibrillation (High Heartbeat and High HRV)
Fever (High body temperature)
state of mind like anxiety, insomnia, etc. (High HR, high HRV, EDA)
Asthma (low RR)
and more...
You can check the hackster.io documentation for the first prototype with nodeMCU, but the following components are what we will be aiming to use for the version we will be making as a part of this prototype -
Using the Arduino Nano 33 IOT instead of NodeMCU with Edge Impulse.
Changing the pulse oximeter from MAX30100 to MAX30102
Using MAX32664D for advanced SPO2, HR, RR, HRV and blood pressure based algorithms
Storing/pushing all data collected on heroku server or edge impulse
MLX90614 medical version (DAA) for more accurate body temperature measurement
SEEED GSR sensor for psychological health monitoring like loss of sleep, anxiety, stress, etc.
Adding medical data of symptoms like arrhythmia, fever and hypoxia in edge impulse to identify a disease instead of just identifying that the vitals are abnormal.
Better and more ergonomic design
Better integration with the NeoSensory buzz to enhance even minute fluctuations
PCB designing and manufacturing
Integration with IOTA - an upcoming virtual currency optimal for IOT devices
How and why each of the above components/technologies are going to be used, will be the agendas for the upcoming blogs. So stay tuned!!
Top Comments