Project proposal by Apoorva Vadhana and Dror Margalit
There is no place I’ve made more mistakes than in front of public places’ trash cans. On my way somewhere, walking by three different trash cans, I need to decide fast: can my soda go in the green bin, or can my paper cup even be recycled?
Trash cans in Central Park
This decision that we are asked to make regarding waste separation is not trivial. Public spaces recycle different materials differently, and they use different color coding to indicate where to dispose of what. Whatsmore, many people do not want to spend time to stop and read long and confusing charts about what to do with their waste. The result is an amount of waste that is too large to grasp. The United Nations estimates that per year, 23-37 million metric tons of waste will be added to the ocean by 2040.
Charts on trash cans in New York City
To make recycling more efficient, we can utilize technology better in two main ways. First, eliminate the choices people need to make when they dispose of trash. By training machine learning models on different materials recycling destinations, we will sort trash at the moment it is being thrown. Second, we will make the process of recycling social. We will visualize how much trash was thrown in the recycling bin and connect it to different bins to compare how other areas waste.
The system will be built on top of three trash cans. Users will place their trash on a platform that will be next to a blank background. Next, we will train an AI model that will scan the item. Using a stepper motor connected to a gear, the item will be moved to the right bin. Last, two servo motors beneath the platforms will be opened, and the item will fall into the bin.
