LitterCaching

Inspiration

Waste management is a growing problem, and recycling is often inefficient due to improper sorting and negligence. When was the last time you saw a can or bottle laying out on the street? Not too long ago, we bet. Thus, we wanted to create a smart, low-cost solution that can identify cans automatically using machine learning and edge computing. Then, users can use an Arduino Uno to control a single servo-powered claw on a rubber-band tensioned PVC pipe contraption for easy deposit into a box that's clippable onto backpacks with additional clips for plastic bags to increase storage and an IR sensor to count cans automatically. Combining all themes of promoting sustainability, wearable tech, and video games by adding gamification such as score counting of cans collected.

What it does

LitterCaching is an intelligent waste-sorting system that uses an ESP32-CAM to identify different types of cans using machine learning. Once detected, servos and a 3D-printed mechanism sort the cans into appropriate bins. The goal is to make recycling more efficient and accessible.

How we built it

We used an ESP32-CAM to capture images of cans, which were processed using Edge Impulse to classify them based on their material or brand. The Arduino Uno controlled the servo, button, and IR sensor for the mechanically heavy side of the project: the picker-upper. The entire system was mounted on a 3D-printed structure designed to hold and direct the cans backward into the deposit box.

Challenges we ran into

Training an accurate machine learning model with limited dataset samples.
Optimizing the ESP32-CAM to capture clear images under varying lighting conditions.
ESP32-CAM COMM port not being recognized or other errors.
Ensuring smooth servo movement without excessive power consumption.
Designing and printing an effective sorting mechanism within the time constraints.

Accomplishments that we're proud of

Successfully integrating Edge Impulse for real-time can detection.
Creating a functional prototype that can grab cans and deposit them quickly.
Overcoming hardware and software compatibility challenges.
Learning 3D modeling and printing to create an efficient sorting system.