EcoBin

This project, known as EcoBin, is an intelligent recycling station specifically designed for the Quick Service Restaurant (QSR) sector to enable a circular economy for packaging. It functions as an automated waste management node that allows packaging manufacturers to recover their specific materials directly from the point of disposal, transforming restaurants into efficient collection hubs. By sorting waste at the source, the system significantly reduces the volume of trash sent to landfills and lowers waste management costs for establishments while ensuring high-quality material recovery.

The system operates through a seamless three-step process that requires zero effort from the consumer, eliminating the need for manual sorting. When a user places an item into the single disposal opening, an integrated computer vision system—utilizing RGB-D cameras and deep learning models like YOLOv8—instantly identifies the object's shape (such as cups, boxes, or lids) and detects potential food contamination. Once identified, a modular rotating mechanism automatically routes the item into one of several internal compartments organized by "form-factor," ensuring that items are stacked efficiently to maximize storage capacity before collection.

The core innovation of EcoBin lies in its shift from sorting by material to sorting by "form-factor," which allows for much higher stacking density and drastically reduces the carbon footprint and costs associated with waste logistics. Furthermore, the project introduces a disruptive "Hardware-as-a-Service" (HaaS) business model where packaging manufacturers fund the technology to meet their Extended Producer Responsibility (EPR) obligations, providing the equipment to restaurants at little to no cost. By using passive Edge AI for classification, the system overcomes the historical barrier of human error in recycling, achieving high adoption rates because it asks nothing of the user while providing high-purity material streams for recyclers.

Moonshot team: Ge Qingyuan, Hemeng Wang, Nai-Chi Cheng, White Daniel Ephram.

Mentor: Carlos Gonzalez De Miguel.