A fun project that introduces you to AI, electronics, and mechanical engineering—no experience necessary! You'll learn how to train a TensorFlow model using Google's Teachable Machine to recognize two different categories of items, then build a machine to sort them.
This makes a great short lesson (1-2 days of class time) for high school-level Engineering or Computer Science courses. The machine can be built using low cost, easy-to-find, reusable parts—perfect for classrooms or homeschoolers.
In addition to paper, tape, and scissors, you'll need:
- A laptop or desktop computer with a webcam (Mac, Windows, or Linux should work)
- A microcontroller compatible with CircuitPython or the Arduino environment such as the Pico 2 W or Adafruit Feather ESP32-S2
- A micro servo, SG90 or similar
- Male-Female Jumper wires to connect the micro servo to the microcontroller
- A USB cable to connect your laptop to the microcontroller
We recommend Adafruit for both tutorials and the components needed to complete the project, but you can source them from a number of places including SparkFun or Amazon in the USA.
See Tiny Sorter Experiment from 2019 for the original project guidelines. See Design, Make, and Teach for up to date instructions for using more modern microcontrollers like the Raspberry Pi Pico and for resources how to incorporate tiny-sorter into your lesson plans!
This project builds on the Tiny Sorter Google Experiment launched in 2019, which combined Teachable Machine with Arduino. It's been updated to modernize the logic for compatibility with newer versions of Chrome and a wider range of microcontroller and development environments.
- arduino: Sketch for installing on a microcontroller using the Arduino environment (C++)
- circuitpython: Source code for installing on a microcontroller using CircuitPython (Python)
- p5js: Source code for the web page to run your model. View it live on p5.js!