ROBbish

Main Section of Site
Information Section of Site
Deep Learning model predicting the category of waste

Inspiration

Los Angeles has implemented a new law called Organics L.A. requiring all residents and businesses to compost their food scraps and separate “green” waste from other trash. Organic waste, which includes food scraps, yard trimmings, and more, makes up a significant portion of waste that goes to disposal in landfills and generates methane gas, a potent greenhouse gas that has negative effects on the environment. Failing to separate green waste correctly could result in fines of up to $500 beginning in 2024. LA residents can face: Fines of up to $500, which could be imposed on those who contaminate their green waste with the wrong items. To recycle incorrect items such as, foam egg cartons, and bubble wrap. We have developed an innovative idea to harness the power of AI in assisting people of all ages and backgrounds to correctly sort waste into appropriate bins. During the hackathon, we observed that items were frequently misplaced in incorrect bins, indicating a lack of understanding of proper waste disposal. Our AI-based solution aims to simplify the waste disposal process and promote sustainable practices. The solution provides users with real-time feedback and guidance, allowing them to dispose of waste correctly and reduce the environmental impact of incorrect waste disposal practices. Robbish is a cutting-edge robot designed to assist in proper waste disposal by guiding users on which bins to use. By utilizing the camera on a user's device, Robbish enables users to take a picture of their waste item and provides real-time information on the appropriate bin for disposal. This innovative solution simplifies the waste disposal process and promotes environmentally friendly practices. It helps individuals who are faced with the confusion of not knowing where to throw away their trash, making waste disposal more accessible and efficient. With ROBbish, users can be confident in their waste disposal choices and contribute to a healthier environment.

What it does

Turn on your phone camera.
Take a picture of the item you’d like to toss.
Let ROBbish guide you to the correct bin

How we built it

We used HTML, CSS, Node.Js, React, java script, and bootstrap for front end design, and functions. We also used Flask for the backend to get the API endpoint. Used a custom pretrained deep learning model to detect the type of waste (organic, recyclable, ..). Deployed this prediction model in the flask. Tried to deploy the flask model in the google cloud. Also, for our presentation we use Canva. We tried to make it as lively as possible to make it entertaining and informative.

Challenges we ran into

Integrating a camera into our website.
Finding a suitable model that precisely classifies the image into different kind of waste.
Deploying flask with deep learning model to google cloud

Accomplishments that we're proud of

Learn to use React efficiently.
Being able to use deep learning to identify and categorize garbage.
Stepping out of our comfort zone.

What we learned

Throughout the event, we experienced moments of both intelligence and inadequacy, as we navigated the process of completing a project within a 24-hour timeframe. There were moments of exhilaration, but also exhaustion as certain aspects of the execution did not go according to plan. However, the experience taught us the importance of coming prepared to a hackathon, including ensuring software is downloaded to the brainstorming of project ideas and necessary software tools. Through collaboration with teammates, we learned to work in a team environment and embrace challenges, even when faced with limited knowledge. Overall, the hackathon allowed us to gain valuable insights into teamwork, time management, and the importance of preparation in project execution.

What's next for ROBbish

Looking to the future, we envision an exciting feature to be added to our solution that incorporates smart trash cans and a pixy cam, enabling users to dispose of their waste items more efficiently. The initial step involves locking the trash cans until a user approaches with an item. The pixy cam will then detect the waste item and unlock the correct bin, streamlining the waste disposal process. This novel solution eliminates the need for users to take pictures of their waste items and wait for instructions from ROBbish, making it even faster and more convenient. These smart trash cans can be implemented in public areas or homes, making waste disposal easier and promoting sustainable practices. Our solution demonstrates our commitment to reducing the environmental impact of incorrect waste disposal, and we are excited about the potential impact this feature could have on waste management practices.

Built With

Submitted to

AthenaHacks 2023

Created by

I worked on the front-end of this project. I used node.js, React.js, HTML/CSS, Bootstrap. I also implemented a little bit of JavaScript. I also attempted to use twilio into our project.

bnguerrero
I take immense pride in the team I collaborated with during the hackathon. Our diverse computer science backgrounds and strong work ethic allowed us to work collaboratively, leveraging our respective strengths to deliver our best until the very end of the 24-hour hackathon. I am equally proud of myself for stepping out of my comfort zone and taking risks, as I am often hesitant to pursue project ideas that I believe to be ambitious. However, the encouragement and support from my teammates gave me the confidence to explore these ideas and pursue them to fruition. The experience taught me the value of teamwork and the importance of embracing challenges, as well as the power of a supportive team in bringing the best out of oneself. Overall, I am proud of the work that my team and I accomplished during the hackathon, and I am grateful for the personal and professional growth that the experience brought me.

Gisela C
I enjoyed collaborating with motivated team members in my first ever in-person hackathon. I worked on the backend of ROBbish, which was a pretty challenging task. We needed to integrate a deep learning model into the API and deploy it on the Google Cloud. Although I had experience working with different backend frameworks like Node.js, Flask, and Django, integrating an ML model was a new challenge for me. Hackathons like these help me to think out of the box and push my boundaries.

NagaHarshita Marupaka
Mariella Galvez Perez