The goal of this project is to enable people who have wearable computing devices and sensors to build interactive, intelligent systems…. even if they have limited computing experience.
While cheap, lightweight sensors have become more and more available (smart watches), their applications have largely been limited to health applications. The average developer doesn’t have the expertise to integrate multiple streams of data, filter the data, and apply machine-learning algorithms in order to extract meaning from the raw signal. Last year, a team of seniors developed FlyLoop, which is a modular java-based framework for developing these kinds of systems. This takes a step in the right direction, but still requires significant programming to construct even the most basic system.
The good news is that FlyLoop is primed for a team to explore better ways to interact with it. I would like someone to create an interface in which it is simple to rapidly develop, save, and share systems that accept physiological signals as input and outputs, for example, an emotional state.
Build an interface that simplifies interaction with FlyLoop. This could be a drag-and-drop application, a browser-based front-end, tangible user interface… feel free to use your imagination.
- Making wearable computing applications more accessible opens the door to an entirely new generation of software that responds not just to our mouse clicks and key presses, but potentially to our cognitive and emotional state.
- Last year, the work from the FlyLoop team was published and presented at an international conference in Germany. I would like to try something similar with this group.
- Finally, I expect to use the platform in my own research for years to come.
- It should be built on top of FlyLoop (simply because it’s probably not worth your time rebuilding an entire new infrastructure). This means that the high-level features available to the user should roughly mirror FlyLoop’s modules – filters, data sources, outputs, learners, calibrators.
- The application must be broadly accessible
- The application must limit hurdles to use (for example, extensive libraries that someone would have to download first).
- Data flows should be savable and sharable.
- Ideally: there is a way to monitor data as it flows between components in the system.
You will be able to access my increasing supply of small sensors to experiment with the platform – eye-tracker, heart-rate monitor, EDA sensor, leap motion, etc.
Open-source software under the MIT License
Point of Contact
Prof. Evan Peck, Department of Computer Science