AI sports data logger

Make an AI sports data logger with micro:bit CreateAI that logs how much time you and others spend running, walking and being still. 

Stap 1: Begrijp het

How does it work?

In this project, you’ll train a machine learning (ML) model to recognise when you’re running, walking and being still.

You’ll combine that model with a MakeCode program that uses the micro:bit’s data logging function to record what action you are doing every second.

This project could be useful in sports such as football or netball where you need to analyse how active certain players are. You could also use it to monitor how much time you have spent running or walking during school breaks or a during a workout.

Wat is machine learning?

Machine learning (ML) is een vorm van kunstmatige intelligentie (AI) waarbij computers kunnen leren en besluiten kunnen nemen op basis van gegevens.

De ML-modellen zijn door mensen getraind om hen te helpen deze beslissingen te nemen, bijvoorbeeld om verschillende 'acties' te herkennen wanneer je je micro:bit op verschillende manieren verplaatst.

Wat moet ik doen?

AI-systemen hebben mensen nodig om ze te ontwerpen, te bouwen, te testen en te gebruiken. Je gaat gegevens verzamelen om een ML-model te trainen, te testen, te verbeteren en te combineren met computercode om een slim apparaat te maken dat AI gebruikt. Je gaat een micro:bit en de micro:bit CreateAI website gebruiken om dit te doen.

Stap 2: Maak het

Benodigdheden

• A micro:bit V2, USB cable, and a battery pack with 2 x AAA batteries
• A computer (e.g. desktop, laptop, or Chromebook) with access to the micro:bit CreateAI website, using a Chrome or Edge web browser
• Als je computer Bluetooth niet ingeschakeld heeft, heb je een extra micro:bit V2 nodig
• A strap and holder, or another way to attach the micro:bit to your wrist (e.g. flexible craft stems or elastic bands)
• You may also find our micro:bit CreateAI teaching tips useful

Verzamel data samples

When you open the project in micro:bit CreateAI, you’ll see we’ve given you some data samples for ‘running’, 'walking’ and ‘still’ actions:

You can add your own movement samples using the micro:bit's movement sensor or accelerometer.

In micro:bit CreateAI, klik op de 'Verbinden' knop om je data collectie micro:bit te koppelen en volg de instructies.

Attach the data collection micro:bit to your left wrist like a watch, with the logo at the top. Click on the first action, ‘walking’, and click 'Record' to record your own data samples. Als je een fout maakt, kun je alle samples verwijderen die je niet wilt. Je kunt ook op knop B op de micro:bit drukken om de opname te starten.

Als je continu wilt opnemen gedurende 10 seconden om 10 samples te krijgen, klik op de drie stippen naast de opnameknop en selecteer die optie.

Now record your own data samples for the ‘running’ action, then the ‘still’ action, making sure for ‘still’ that you collect samples in different positions, such as facing up and down.

Train en test het model

Klik op de knop 'Train model' om het model te trainen en vervolgens te testen. Try walking and see if ‘walking’ is the estimated action, then running to see if ‘running’ is the estimated action. Keep still and see if ‘still’ is estimated. Give your micro:bit to someone else to wear (making sure they put it on the same wrist and in the same orientation) and see if it works as well for them.

Verbeter je model

De meeste modellen kunnen worden verbeterd met meer gegevens. Als het model verbeterd moet worden, klik dan op ‘← Bewerk data samples’.

You can delete any data samples which you think don’t fit or add more samples from yourself and other people.

Train het model opnieuw en test het opnieuw.

Plaats het model en de code op je micro:bit

Klik in micro:bit CreateAI op 'Bewerk in MakeCode' om de project code te zien in de MakeCode editor.

Je kunt de code wijzigen of het gewoon uitproberen zoals het nu is. Attach your micro:bit using a USB cable, click on the ‘Download’ button in the MakeCode screen, and follow the instructions to transfer your AI model and the code blocks to it.

Attach a battery pack to the micro:bit and put it on, ready to test out.

Collect your data

First press buttons A and B together to delete any old data logs from your micro:bit. Reset the timer by pressing the reset button on the back of the micro:bit. Press button A to start logging and button B to stop logging. 

Your data will stay on your micro:bit even if you disconnect the battery or press the reset button.

Analyseer je gegevens

Koppel het batterijpakket los en sluit de micro:bit weer aan op een computer. De micro:bit verschijnt als een USB schijf genaamd MICROBIT. Kijk in het MICROBIT station en open het MIJN_DATA bestand om een tabel met je gegevens in een webbrowser te zien:

The time stamps in the log represent the amount of time that has passed since your micro:bit was powered on or reset.

Click on Visual preview to see a graph of your data:

You can also click on the Copy button and then paste your data into a spreadsheet.

Hoe de codeblokken werken

This program uses a variable called ‘logging’. A variable in a computer program is a container for storing data which can be accessed and updated while the program is running. In this program, the variable ‘logging’ controls if the micro:bit is logging or not and can be set to ‘true’ or ‘false’. Variables that can be set to these two values are called ‘Boolean’ variables.

When the program starts, the variable ‘logging’ is set to false. A show icon block is used to display a ‘no’ icon on the LED display to indicate the micro:bit is not logging. The set columns and set timestamp blocks create labels for the data logging table your micro:bit will produce.

The on button A pressed block is used to set logging to ‘true’ and show a ‘yes’ icon on the LED display. The on button B pressed block is used to set logging to ‘false’ and show a ‘no’ icon on the LED display. And the on buttons A + B pressed block sets logging to ‘false’, displays a skull icon, and deletes any log.

Finally, an ‘every’ block is used to check every 1,000 milliseconds or second if the micro:bit is logging. If it is, an ‘if then else’ block is used with ‘is ML detected’ and ‘log data’ blocks to record a 0 if you are still, a 1 if you are walking and a 2 if you are running in your data logging table. If the micro:bit cannot detect what you are doing, it records a -1 in the table. Bigger numbers are used for more active actions, so the resulting data logging graph gives you a clear visual record of how active you have been.

Beoordeling

How accurate is the AI sports logger in tracking your movements? How could you improve its accuracy? Who would find this device particularly useful? How does it compare to the Step counter or the Movement data logger projects?

Stap 3: Breid het uit

• Add a fourth action such as ‘throwing’ for sports like netball or tennis.

• Add up how much time you spent on each activity. You could do this saving your data as a CSV file, opening it in a spreadsheet and using a formula such as =COUNTIF(B2:B70,1) Where B2:B70 is the range of the activity cells, and 1 is the activity number meaning ‘walking’.