Building a Magic Wand with TensorFlow Lite Micro - Tutorial

Learn how to build a gesture-recognizing magic wand using an Arduino Nano BLE Sense 33 in this comprehensive tutorial from the tinyML Summit 2021. Explore the process of gathering data, training models, and deploying IMU-based models for gesture recognition. Gain hands-on experience with the Arduino IDE and Colab scripts while understanding the intricacies of feature generation for raw accelerometer and gyroscope data. Dive into practical application development using TensorFlow Lite Micro and Arm's CMSIS-NN library. Follow along with topics including Arduino IDE installation, board connection, ESP32 MCU usage, Bluetooth connectivity, model creation and sizing, quantization accuracy, and custom gesture implementation. For optimal learning, consider acquiring the Arduino TinyML Kit to participate fully in this virtual hands-on experience.

Introduction
Overview
Requirements
Shortcut
Closeup
Installing Arduino IDE
Why Arduino
Board Connection
ESP32 MCU
Upload Magic Wand Example
Bluetooth Connection
Collab
Model
Model Size
Model Upload
Quantization Accuracy
Balancing Network and Accuracy
Testing
Custom Gestures