The opioid crisis has hit neighborhoods across the country. Deaths have increased from 8,400 in 2000 to over 28,000 in 2014. We try to tackle this nationwide crisis by proposing PEARL, a wearable wireless pulse oximeter on a person's earlobe. PEARL transmits real time data to a cloud (possibly Amazon Cloud) where the data is processed and the person's heart rate, breathing rate, and SpO2 content is calculated. During an overdose, breathing becomes heavy and labored and SpO2 content decreases to unviable levels. With PEARL, once these levels fall below a threshold, a signal will be sent to local first aid responders through the cloud. We hope that this will decrease the response time for these individuals to receive life saving treatment. PEARL is discrete, mobile, and easy to use. This device has a wide reach - it can not only alert first responders to individuals who overdose but also for patients with chronic health issues or expecting mothers with a drug addiction.

What it does

PEARL is an earring that doubles as a pulse oximeter. It transmits real time data to a cloud that processes this information and alerts first responders if there is an anomaly (when SpO2 falls below a threshold). This discrete device repackages existing and reliable pulse ox technology into a wearable medical 'safety net' that will hopefully allow first aid to assist patients who have overdosed in a timely manner.

How we built it

We build a rudimentary pulse oximeter with an arduino board and a phototransistor. The light source is supplied by an iPhone camera with a red film over it. The phototransistor is placed on one side of an individual's ear and the iPhone light is placed at the other side. This data displayed in the Arduino and it is exported into a file that can be read by MATLAB.
Our MATLAB code visually displays the data and calculates the SpO2 content. Due to limitations in the hardware, we were unable to collect viable data to analyze. We used a dummy data set to test if our code works. The code has a low pass filter that removes high frequency noise associated with data collection. We were also able to extract heart rate and breaths per minute from this data although SpO2 content was the only metric used for the threshold alert system. Once the average SpO2 content falls below a threshold, an alert will be sent from the cloud to local emergency responders. We were unable to program this into MATLAB. For now, if the SpO2 content falls below a threshold, MATLAB plays a sound as a warning.

Challenges we ran into

None of us knew how to code in any language except for MATLAB (which may people and professionals have told us, isn't really a coding language). Although we knew how a pulse oximeter works, we had limited hardware to build this device. We worked with arduinos for the first time, and we had many issues coding and downloading the program onto the board. Troubleshooting and trying to understand these issues took the majority of the first night and day. We also ran into a few challenges with collecting viable data for our project. The signal resolution from our device was not high enough to pick up on the minute changes in absorption from the pulsating vessels in our subjects ear. Because we were limited in our hardware, we decided to create a dummy vector of data to simulate normal and overdose pulse oximetry signals.

Accomplishments that we're proud of

This weekend we had a crash course on arduinos, coding in other languages, and working with some circuits. We befriended some brilliant and kind hackers who generously offered up their time to answer all of our questions and help us setup our arduino board. They were extremely excited when teaching us these new concepts and their enthusiasm definitely rubbed off on us.

On a completely unrelated note, those two kind hackers also helped us on our op-amps homework assignment that is due on Monday. Although we only understand it marginally better, it is still a huge accomplishment in our books!

What we learned

We learned how to code some arduino boards and how to assemble arduinos and sensors. We also learned (through trial and error) the best way to use sensors and collect data. Through talking to other hackers, we received some advice on how to apply to grad schools and some useful websites on how to learn how to develop apps and learn C or C++. This experience has made us excited to join more hackathons (and pay more attention to our biomedical instrumentation classes).

What's next for Pulse Oximeter for Ear Lobes (PEARL)

Because we did not have a breadboard, we were unable to connect a infrared light and detector. We hope to incorporate an IR light in addition to our red light to collect accurate SpO2 data. Many improvements can be made to the hardware to improve it's data collection and decrease the cost of production. We wish to be able to distribute this product to communities through public service centers such as libraries, planned parenthood, convention centers, etc. Our software is currently based in MATLAB. To make it more compatible with other devices/softwares, it should be translated to javascript or python. Another big step would be to get into contact with Amazon web services and link the cloud with our device. Information about this product can then be disseminated to first aid responders/police department across the country.

Built With

Share this project: