Learning about Paul Kotler's challenges with anxiety inspired us to think about how technology could help him and better track and monitor his anxiety. Our hope is that Tranquil Tracker will help Paul and others recognize their anxiety before they begin to experience it so that they can more easily take preventative measures.
How it works
Tranquil Tracker combines physiological data with self-reported online data to develop a user's anxiety risk profile which predicts the user's likelihood of experiencing an anxiety-driven neurological disruption. Time-stamped physiological data (heart rate, body temperature and skin conductance) and self-reported data (pending appointments, geo-location, audio cues from the user's phone's speakers, etc) are trained on a neural network machine learning process to assess what combinations of metrics signal:
- an increase in anxiety
- an impending episode
- an neurologically disruptive episode such as a panic attack
Each of these signals can be linked to trigger an appropriate technological response such as:
- playing soothing music
- controlling environmental conditions in a Smart home
- notifying a caregiver via SMS or email
- flashing a wearable LED to alert the user and others of their changing mood
The user's profile is saved to periodically re-train the prediction metrics. The user can access their time-stamped data at any time for review allowing them to better understand the triggers and causes of their anxiety. The user can also modify their profile by adding or removing self-reported data after-the fact and by logging high-anxiety periods that occurred while the user's physiological readings were not being recorded. Any modifications to the user's profile will prompt the neural network to re-evaluate the data and develop new metrics for tracking the user's anxiety.
Challenges I ran into
Collecting physiological data and extracting meaning from it turned out to be much trickier than expected.
Accomplishments that I'm proud of
Very proud of the conceptual framework, methodology as well as the potential Tranquil Tracker has to improve the lives of anyone suffering from issues relating to anxiety.
What I learned
Anxiety can overwhelm a person's ability to communicate and cause significant distress. For an autistic person this is especially challenging because of their increased sensitivity and existing communication difficulties. Speaking with Paul and working on this challenge really drove home an appreciation for how anxiety and autism can mask an intelligent, vibrant human being. Another lesson learned was that collecting collecting biological signals data in real-time is a complicated matter. Adapting high-quality existing hardware can sometimes be more practical than utilizing hobbyist-grade parts because noisey data can be a significant hindrance - especially when you're simultaneously learning how to analyze that data type.
What's next for Tranquil Tracker
Continue developing the prototype: a soft fiber-optic fabric wearable with embedded biosensors, bluetooth connectivity and LED display; embedded speakers optional.
Log in or sign up for Devpost to join the conversation.