Inspiration
We as a team were inspired by the fly brain advancements that have been happening in recent times. We believe that just like the developments in AI led to the tech world we live in today, these developments could very well be the start of the next great era of tech. Fruit flies can react to, and escape from, flyswatters in milliseconds. We figured, since this brain system is fully mapped, why not utilize it to help improve bikers reactions to possibly fatal accidents.
What it does
Our program utilizes a camera, and a raspberry pi in order to record live footage. It takes in the video footage and scans it for any activity. The fly brain is scanning for objects rapidly approaching, and at the same time, the program is trained with the Waymo public database to help detect whether those objects are vehicles, or something else. If it is a vehicle then the raspberry pie is signaled and triggers sounds and LEDs
How we built it
We build this utilizing the publicly available fly brain released by Google research, in partnership with the publicly available Waymo database in order to identify vehicles and alert the system when they are rapidly approaching. We split the programming into four main parts, those being the FlyEye (inputs from the camera), FlyMind (logic switches), the Hardware outputs , and the Waymo data/samples. We split these roles up amongst ourselves and then combined them all in the GitHub to make our finished product.
Challenges we ran into
The whole idea of utilizing the real neural networks of a fly was very new and difficult for all of us. We spent the majority of our time trying to understand, and troubleshoot how to properly understand the neural networking pathways of the flies. Also with some of our members not having any experience, we spent a good bit of time troubleshooting and setting up their work environments and GitHubs.
Accomplishments that we're proud of
We are very proud of the final product we created. We truly believe that the human brain is one of the most powerful computers in existence, and we believe that these developments are putting us just one step closer to understanding it a little better.
What we learned
Personally, with this being my first hackathon I learned an absolute ton. This whole experience and opportunity has thought me so much about the whole development process. Before ShellHacks, I didn't know the first thing about GitHub, collaborative programming, or fly brain. But even the more experienced members of our team learned a lot at ShellHacks. This was all of our first time using hardware, so the raspberry pi proved to be a very steep learning curve for us as well.
What's next for I Fly
Now that we have a basic system for giving the fly a "body" with the raspberry pi, and we have built and proven a use case with vehicle detection, we are excited to modify the fly to be used for a whole array of different purposes.
Log in or sign up for Devpost to join the conversation.