TinyML with Arduino Nano RP2040 Connect

如您所知，TinyML 领域正在不断发展壮大。因此，请记住，在不久的将来，将包含更多带有实践项目想法的理论解释部分。

针对电池供电设备的微型机器学习被广泛定义为快速扩展的机器学习技术和应用领域，包括硬件（专用集成电路）、算法和软件，能够以极低的速度执行设备上传感器数据分析功率，通常在 mW 范围内及以下。它消除了将数据发送到云进行分类的要求，从而提供了更高的安全性。此外，耗电量大的处理器正在被微型 MCU 所取代。当然，也有限制。这些限制来自有限的硬件资源、时钟速度等。不过，仍有一些应用领域不需要高计算量，并且需要基于机器学习的解决方案。在这种情况下，TinyML 就会出现。它可用于检测工厂中的机械异常，可以预测仪器、医疗保健领域等的维护需求。TinyML 应用领域广阔，前景广阔。

本课程的主要目标是从数据收集、模型训练、测试和部署开始熟悉 TinyML 开发。本课程使用具有 265KB RAM 和 16MB 闪存的低成本 Arduino nano RP2040 连接板，内置加速度计、陀螺仪、麦克风、温度传感器和无线连接模块（WiFi+蓝牙），此处演示的所有示例均在此测试木板。

本课程适合
初学者，有兴趣在低成本、低功耗微控制器中开发机器学习模型

Published 10/2022
MP4 | Video: h264, 1280×720 | Audio: AAC, 44.1 KHz, 2 Ch
Genre: eLearning | Language: English | Duration: 16 lectures (1h 35m) | Size: 669 MB

Machine learning model development for tiny low power microcontroller such as Arduino nano RP2040 connect

What you’ll learn
To be able to understand hardware requirement for development of machine learning model for tiny MCUs
Understanding the tinyML development framework
To be able to create tinyML projects based upon hand gesture
To be able to develop tinyML model with audio keyword detection

Requirements
Arduino nano-RP2040 connect board, USB cable, PC/Laptop, Basic knowledge about Arduino IDE, basic knowledge of machine learning, basics of embedded C/C++

Description
**Note: This course is not finalized yet. As you know, the TinyML field is constantly growing and developing. So, keeping in mind more sections with theoretical explanations with hands-on project ideas will be included in the near future.

Tiny machine learning, which targets battery-operated devices, is broadly defined as a rapidly expanding field of machine learning technologies and applications that includes hardware (dedicated integrated circuits), algorithms, and software that can perform on-device sensor data analytics at extremely low power, typically in the mW range and below. It eliminates the requirement to send data to the cloud for classification thus providing more security. Also, power-hungry processors are being replaced by a tiny MCU. Of course, there are limitations. The limitations came from limited hardware resources, clock speed, etc. Still, there are several application areas where high computation is not required and a machine learning-based solution is desirable. In that case, TinyML will come into the picture. It can be used to detect anomalies in machinery in a factory, it can predict maintenance requirements of the instruments, healthcare field, and so on. The application domain of TinyML is wide and the future is bright.

The primary objective of this course is to be familiar with TinyML development starting from data collection, model training, testing, and deployment. A low-cost Arduino nano RP2040 connect board having 265KB RAM and 16MB flash with in built accelerometer, Gyroscope, Microphone, temperature sensor, and wireless connectivity module (WiFi+Bluetooth) is used in this course and all example demonstrated here is tested on this board.

Who this course is for
Beginner, interested to develop machine learning model in low cost, low power microcontroller