Inspiration
Inspired by the growing waste crisis in world, and observing improper disposal practices around us, we wanted to work on a project that would contribute to solving this problem one bin at a time.
What it does
Flip is a smart waste management system that combines both AI and Robotics to sort waste into recyclable and non-recyclable categories.
How we built it
Flip is built from the ground up with sustainability in mind, being composed of fully recycled materials. We utilize a raspberry-pi and arduino to handle control for the device as well as communication with our react website that boasts the usage statistics for our device. We leverage Google's gemini to handle object recognition and pass this information onto our onboard electronics that handle the sorting.
Challenges we ran into
Flip is a device that leverages multiple technologies, each carrying their own unique challenges. Arduino and Raspberry-Pi were a wonderful asset to Flip, however relaying information gathered by AI and translating that to movements by Flip proved to be challenging. Furthermore, handling a continuous camera feed and taking snapshots to process with AI was more complex a task than it seemed at first glance. Flitering noise in the camera feed as well as ensuring object recognition remained accurate took many trials, testing with various objects and lighting conditions.
Accomplishments that we're proud of
We are proud of how our project has come together, each component coming together to form a cohesive product haas been both a satisfying and rewaring journey.
What we learned
We have learned quite a bit on what it takes to bring such a concept to life. Combining AI through Gemini, Vision, Raspberry Pi, Arduino, and MongoDB has proved to be quite a powerful combination of technologies with great capabilities for usage. The things we have learned about these technologies will be a great asset in improving this project as well as building future ones.
What's next for Flip
Flip is a very early stage, as an early prototype flip still has great room for improvement and refinemnets. The first step being a more refined chassis as well as a more polished AI integration.
Log in or sign up for Devpost to join the conversation.