Virtual Booth

Sorting systems are used to categorise products for appropriate packaging, offering quick and labour-saving sorting compared to manual methods. As industrial-grade sorting systems are typically large-scale, citizens oftentime lose out on this useful technology.

Our project aims to make this technology accessible for everyday use aligning with SDG 9: Industry, Innovation, and Infrastructure. Usually, sorting systems are set to only sort by size, weight and colour. Our solution is an indoor, beginner-friendly automatic sorting system that can sort objects based on these characteristics and more using machine learning. In order to demonstrate the capabilities and the simplicity of using the system, we present it as a toy sorter.

The system comprises of three main parts: 1) Customising categories : Using TeachableMachine, users can train the system by uploading images of different categories. The system then sorts objects using webcam footage. 2) Sensing: A conveyor belt powered by a dc motor, with an infrared sensor at the start to detect objects and a webcam above. 3) Sorting: A Ferris wheel-style revolving cabinet with detachable, lined compartments to prevent damage. The cabinet rotates based on the sequence and category of objects, controlled by Raspberry Pi and a stepper motor. The unique vertical arrangement saves horizontal space, making it suitable for smaller environments.

In short, our system offers a versatile solution for small-scale applications. Its flexibility allows for one-time purchases to serve various purposes, such as family businesses.

Keywords: Automatic Sorting System, Small-scale, SDG 9: Industry, Innovation and Infrastructure, AI Machine Learning, Customisable Sorting, User-friendly Interface, Revolving Cabinet