Squats Counter Using TensorFlow Lite and Tiny Motion Trainer

Published Sep 07, 2021
 2 hours to build
 Intermediate

Squats counter that can detect squats by measuring the accelerometer readings and prints it on the Serial monitor

display image

Components Used

ARDUINO NANO 33 BLE SENSE
Bluetooth Development Tools - 802.15.1 ARDUINO NANO 33 BLE SENSE WITH HEADERS MOUNTED
1
Micro USB Cable
USB Cables / IEEE 1394 Cables 3FT STRONG MCROUSB-B/USBACBL
1
Description

Squats-Counter

This 'Arduino Nano 33 BLE Sense' based Squats counter can count the number of Squats performed by an individual using the Accelerometer readings and TinyML based Machine learning model.

 

Demonstration

The Squats counter can detect squats by measuring the accelerometer readings and prints them on the Serial monitor.
The device is to be attached to the individual's thigh and can be powered using a power bank or a battery.

Things used in this project:

I am using this complementary Google IO kit from Sparkfun that includes the Arduino Nano 33 BLE Sense that is capable of running TinyML projects with ease.

 

Steps to follow

1. Upload the tf4micro-motion-kit code to the Arduino Nano 33 BLE Sense

 2. After uploading the code, open the Tiny Motion Trainer Experiment by Google and pair the Bluetooth device with the laptop.

3. Select the appropriate settings for training the model

4. Capture the data required to train your model

5. Train your model

6. Test your model

7. Download the Arduino code and upload it to the Arduino BLE Sense Board

8. Test the model by using the Serial monitor

9. Update the code as per your requirements or use mine: BLE_Sense_Arduino_Code

10. Place the device on your thigh using a Hook and Loop fastener (Velcro tape)

11. Done!

 

 

Codes
Comments
Ad