Let's improve our program with AI
Let's improve our program with AI
Embedded Files
We are going to
create a machine learning model. You may have done this before with images, distinguishing a dog from a cat. Machine learning (ML) is a kind of artificial intelligence (AI) where computers can learn from data and make decisions based on it.
then training it,
and at least using it in the code we will download to the micro:bit.
The tricky part ! first, we have to download a communication program on each boards so that they can communicate via radio link to collect the data
The tricky part ! first, we have to download a communication program on each boards so that they can communicate via radio link to collect the data
We need two boards because one (the black one) is used as the bluetooth connector, since the school computers don’t have bluetooth.
And we start with the micro:bit n°1, the blue one.
We are going to do this first exercise together so you understand how it works and can complete the other exercises on your own.
In this project, the aim is to train a machine learning (ML) model to detect whether we are exercising or not.
To access the exercice click here on this link / then find the image above and click on "open in micro:bit create AI" / after on "start session"-
After a first try you can improve your machine learning, step by step :
1) Click en EXERCISING and change the machine learning model to recognize a specific movement like waving, claping or raising your hand. Click on RECORD for 10 SECONDS
2) Click on ADD ACTION to have 3 different actions
3) Click on TRAIN MODEL
4) Click on EDIT IN MAKE CODE / DOWNLOAD the program in the blue card. Your device is now autonomous.
Exercise 2 AI activity timer : it recognizes two different movements and displays the time spent on each activity.
Exercise 2 AI activity timer : it recognizes two different movements and displays the time spent on each activity.
Exercice 4 AI Light switch : this one is the hardest because of the cards manipulations. The aim is that the movement recognized on the blue card turns on the leds on the black card.
Exercice 4 AI Light switch : this one is the hardest because of the cards manipulations. The aim is that the movement recognized on the blue card turns on the leds on the black card.
First, you train the machine learning model as you did for the other exercises.
When your machine learning model is ready, check on the computer that the program does what you want : the lights on the black card turn on when you clap and turn off when you wave.
Replace the radio communication number 52 with a number from 0 to 255. Each group must have its own number. This number will allow your blue card to send messages to your black card.
- then download the program in the blue card AND in the black card (Pay attention, leave the black card plugged into the computer or use a second battery) and test your project.
Tests des programmes du site microbit.pdf
Page updated
Google Sites
Report abuse