Subway Sweepers

Inspiration

Subway Sweepers was born from a clear need: improving how urban environments handle waste. We observed that many public spaces suffer from inefficient cleaning methods, leading to persistent litter problems. Our solution is a self-contained robot that detects, collects, and properly disposes of trash. By leveraging a Raspberry Pi running a YOLO-based object detection system and an Arduino for precise control of its movement and collection mechanism, Subway Sweepers offers a practical, solar-powered approach to keeping public areas cleaner and more efficient.

What it does

Subway Sweepers is a solar-powered sustainability robot designed to detect and collect trash in real time. Equipped with a Raspberry Pi running YOLO-based computer vision, the robot identifies a variety of waste items in its environment. Once trash is detected, an Arduino navigation + precise scooping mechanism—guided by distance sensors—collects and deposits the waste in the proper disposal area. Whether operating on busy city streets or in subway stations, Subway Sweepers brings high-tech cleanliness to any setting.

How we built it

Our project integrates both hardware and software innovations to bring Subway Sweepers to life. The system’s core is a Raspberry Pi connected to an LCD monitor and a webcam, which works in tandem with a trained YOLO model for object detection. An Arduino board processes sensor data to translate visual inputs into navigation commands, ensuring the robot can autonomously move and clean. The entire setup is sustainably powered by solar panels, while robust rotary motors and a repurposed can plow enhance its operational effectiveness. Additionally, we built an interactive user interface using Streamlit, complete with a generative AI chatbot to provide creative insights and product information.

Challenges we ran into

We initially faced connectivity issues between the Raspberry Pi and our laptops, which forced us to rethink our approach and code through a LCD display. Mechanical challenges with the rotary motors and plow mechanism also required extensive troubleshooting and iterative design improvements, including resoldering and getting new equipment alternatives. Despite these setbacks, each challenge served as a learning opportunity, encouraging us to refine our integration of hardware and software for a more robust solution.

Accomplishments that we're proud of

We're proud that we successfully integrated a YOLO-based object detection system on a Raspberry Pi, enabling real-time trash detection and collection. We're also happy to have figured out an Arduino-driven control system manages sensor inputs and navigation commands, ensuring precise autonomous operation. We're proud of the robot’s hardware—including rotary motors and a plow mechanism—as well as including solar panels for sustainable power. Additionally, we think its pretty cool that we developed a Streamlit-based web interface with an integrated generative AI chatbot to display system analytics and facilitate user interaction.

What we learned

This project provided practical experience in deploying machine learning models on resource-constrained hardware. We refined techniques for integrating YOLO object detection with Raspberry Pi and interfacing sensor data through an Arduino. Key lessons included optimizing energy management using solar power, establishing reliable communication between heterogeneous hardware components, and implementing modular design practices for efficient prototyping and iterative system improvements.

What's next for Subway Sweepers

Enhanced Data Collection: Integrate more sensors to track environmental metrics (e.g., waste levels, graffiti, infrastructure issues) and display them on an interactive dashboard. Robust Hardware Improvements: Upgrade to stronger motors, sturdier plows, and improved chassis design for durability in harsh urban environments. Aesthetic and Functional Upgrades: Refine design for better visual appeal and user interaction. Partnerships and Pilot Programs: Collaborate with local governments, environmental organizations, and private enterprises for real-world testing and deployment. Scalable Deployment: Develop a fleet management system to coordinate multiple robots for broader urban coverage. Community Engagement: Launch outreach programs to educate communities on sustainable practices and the benefits of smart city technologies.

Please note that the updated video is in our deployed website :)