There is a large sub-population of people that struggle with stress, especially stress that compounds to excessive levels. Many do not have the means of communicating this to those that can help, such as the severely autistic or special needs people. Not being able to address the stress until it manifests as a violent outburst or break down is a problem for many caregivers and guardians. Being able to monitor someone's relative stress level would give these people a tool to pre-emptively help, and minimize the damaging effects of compounding stress. Given the target population, we wanted to make something that was easy and convenient to use, something that can be slipped on to instantly start monitoring the individual. Furthermore, we developed the system to easily add more data sources and metrics, allowing it to contribute to the quantified self.
What it does
The device is based on a home-brew galvanic skin response (GSR) sensor. Measuring GSR gives an easy, indirect method of gauging relative stress levels by peeking at the sympathetic nervous system. When stressed, the GSR naturally increases, and vice versa. The sensor reports data to a Raspberry Pi, which is uploaded to a hosted database. An easy to use web-app allows someone to monitor the person from anywhere, and to watch for any irregularities. If abnormal stress levels are observed, the guardian can respond appropriately.
How We built it
The GSR sensor consists of 2 electrodes integrated into an elastic wrist-band. The wrist-band, when worn, contacts the forearm on the top and bottom, providing a comfortable location to measure GSR. A unique Super-Mario “M” allows the positioning to be fun and easy. The electrodes are connected to a simple circuit to measure the GSR, and a 16-bit ADC reads the value. The digitized information is read in by the Raspberry Pi and uploaded to a hosted Mongo database.
Challenges I ran into
Among the challenges we ran into were reading analog signals with the raspberry pi. Originally, we used an Arduino Uno to read the voltages and send them to the Pi via USB. Then BookHolders came along and provided us with an analog to digital converter with an I2C interface; which was fortunate because reading data via USB was unstable and prone to crashing our program. We were also provided with a OLED I2C screen, which allowed us to consider giving extra information about the data.
Accomplishments that I'm proud of
Things that we are particularly proud of are our sensor and its surprising functionality and our ability to display the information on both the small screen on the pi, the android app as well as a live stream and static graph in the browser, allowing the information to be accessed from nearly anywhere.
What I learned
What we learned was that the I2C interface is very easy to use and provides a great deal of efficiency in terms of connections and use of space.
What's next for Stressbuster
We hope to continue developing Stressbuster to more useful and more accurate. In addition to the GSR sensor, we want to include a heart rate sensor in the wrist-band, increasing the accuracy of the sensor. Furthermore, we want to include an “emergency contact” feature, so that appropriate parties can be notified of higher stress levels.