We have all seen the pictures. The images of beaches covered in trash, stacked end over end to crate a massive stack of waste. In these piles, you can see bottles and boxes, cans and cardboard, all prime recyclable materials. But these materials will not be saved. Eventually, giant bulldozers will come and take all of that litter to a landfill, moving the problem out of sight. But what of the materials? The items that could have been reused or recycled? They, along with the rest of the trash, find their way into an incinerator or stack up in a landfill. This accounts for a massive loss of resources each year and contributes to the growing waste management crisis in these areas. But what if we could change that? What if there was a robot that could identify and collect these recyclable items when they were first littered? To answer that question, we created GARI, an easily scalable and affordable solution.
What it does
The Green Automated Recycling Initiative (GARI) uses an array of sensors to navigate an area and locate trash, which is identified by a top-mounted camera and classified into material categories. Once identified as recyclable, GARI is equipped with a claw to capture and transport the soon to be repurposed material back to where it can be recycled.
How I built it
GARI functions using a Google Cloud Vision model that, combined with OpenCV, classifies images of trash automatically based on their main material. On the back-end, we used custom scripts to interface with the APIs and services that make the image recognition possible. We used several Arduinos to operate the mechanical arm and steer the robot towards its objectives.
Challenges I ran into
Integrating the various components needed to make GARI a complete system took quite a while, as it required several Arduinos, Google Cloud, Python, C++, and quite a bit of soldering. The pieces on their own were manageable, but ensuring that each part could interface with the others was far more challenging.
Accomplishments that I'm proud of
Our team worked together very well, helping each other, producing quality work, and adapting to challenges quickly. We learned new technologies and used new tactics to solve our problems. But more importantly than that, our team feels that we were actually able to make an impact with GARI. Completing a project is definitely something to be proud of, however, we are even more proud that GARI can make a real impact in the world by cleaning up the environment and promoting recycling. The potential to help create a better community and a better world through robotics is exciting, and we hope that GARI can help do that.
What I learned
I personally learned a lot about Google Cloud's services and the machine learning methods it offers. My teammates helped each other quite frequently and showcased unique findings of their tasks, raising our awareness of new technologies available and teaching us new ways to solve problems. More importantly, however, we learned that the risk and pressure of taking on a large project just make the completion more rewarding. Being able to create something that makes an impact is an amazing feeling, and our team hopes to continue doing this in the future.
What's next for Green Automated Recycling Initiative (GARI)
Specifically, we hope to deploy GARI at our own campus at Illinois Wesleyan University to help clean up any stray trash and contribute to our already strong campus recycling program. Hopefully, we will use more of our time to improve different aspects of the system to fit even more types of recycling and operations. One of the ideas we had is that we could aggregate the data that GARI produces to find out which places have the largest amounts of litter, and then to apply that data by placing trash cans or recycling bins.