Appendix A: Tools

This is a non-exhaustive list of tools and frameworks that are available for embedded AI development.

A.1 Hardware Kits

A.1.1 Microcontrollers and Development Boards

No	Hardware	Processor	Features	TinyML Compatibility
1	Arduino Nano 33 BLE Sense	ARM Cortex-M4	Onboard sensors, Bluetooth connectivity	TensorFlow Lite Micro
2	Raspberry Pi Pico	Dual-core Arm Cortex-M0+	Low-cost, large community support	TensorFlow Lite Micro
3	SparkFun Edge	Ambiq Apollo3 Blue	Ultra-low power consumption, onboard microphone	TensorFlow Lite Micro
4	Adafruit EdgeBadge	ATSAMD51 32-bit Cortex M4	Compact size, integrated display and microphone	TensorFlow Lite Micro
5	Google Coral Development Board	NXP i.MX 8M SOC (quad Cortex-A53, Cortex-M4F)	Edge TPU, Wi-Fi, Bluetooth	TensorFlow Lite for Coral
6	STM32 Discovery Kits	Various (e.g., STM32F7, STM32H7)	Different configurations, Cube.AI software support	STM32Cube.AI
7	Arduino Nicla Vision	STM32H747AII6 Dual Arm® Cortex® M7/M4	Integrated camera, low power, compact design	TensorFlow Lite Micro
8	Arduino Nicla Sense ME	64 MHz Arm® Cortex M4 (nRF52832)	Multi-sensor platform, environment sensing, BLE, Wi-Fi	TensorFlow Lite Micro

A.2 Software Tools

A.2.1 Machine Learning Frameworks

No	Machine Learning Framework	Description	Use Cases
1	TensorFlow Lite	Lightweight library for running machine learning models on constrained devices	Image recognition, voice commands, anomaly detection
2	Edge Impulse	A platform providing tools for creating machine learning models optimized for edge devices	Data collection, model training, deployment on tiny devices
3	ONNX Runtime	A performance-optimized engine for running ONNX models, fine-tuned for edge devices	Cross-platform deployment of machine learning models

A.2.2 Libraries and APIs

No	Library/API	Description	Use Cases
1	CMSIS-NN	A collection of efficient neural network kernels optimized for Cortex-M processors	Embedded vision and AI applications
2	ARM NN	An inference engine for CPUs, GPUs, and NPUs, enabling the translation of neural network frameworks	Accelerating machine learning model inference on ARM-based devices

A.3 IDEs and Development Environments

No	IDE/Development Environment	Description	Features
1	PlatformIO	An open-source ecosystem for IoT development catering to various boards & platforms	Cross-platform build system, continuous testing, firmware updates
2	Eclipse Embedded CDT	A plugin for Eclipse facilitating embedded systems development	Supports various compilers and debuggers, integrates with popular build tools
3	Arduino IDE	Official development environment for Arduino supporting various boards & languages	User-friendly interface, large community support, extensive library collection
4	Mbed Studio	ARM’s IDE for developing robust embedded software with Mbed OS	Integrated debugger, Mbed OS integration, version control support
5	Segger Embedded Studio	A powerful IDE for ARM microcontrollers supporting a wide range of development boards	Advanced code editor, project management, debugging capabilities