APPLICATION OF MACHINE LEARNING MODEL TO MICROCONTROLLERS - AUTOMATION OF IOT EDGE DEVICES

Author Name: 1. MSc. Vo Hung Cuong 2. MSc. Dinh Thi My Hanh 3. Mr. Tran Cong Danh

Volume/Issue: 02/07

Country: VietNam

DOI NO.: 08.2020-25662434 DOI Link: https://www.doi-ds.org/doilink/12.2021-99385823/UIJIR

Affiliation:

1. Lecturer Of Vietnam Korea University of Information and Communication Technology – The University of Danang
2. The University of Danang, PhD student of Hanoi University of Science and Technology, Vietnam
3. Student, Vietnam Korea University of Information and Communication Technology – The University of Danang

ABSTRACT

The Internet of Things has advanced at a breakneck pace in recent years. As a result, cloud servers are storing billions of records, causing delays for some IoT systems, which must transport data from many devices to the server and execute machine learning computations. As a result of the rapid growth of microcontrollers, a new idea known as edge computing was formed. Tensorflow lite is a big library that allows microcontrollers to employ machine learning models. In this post, we'll develop a system that uses a machine learning model placed on the ESP32 microcontroller to autonomously control lights and fans based on sensors in the surroundings. The Arduino Integrated Development Environment is utilized with TensorFlow Lite for Microcontrollers. With a varied number of neurons, neural networks with two hidden layers are employed.

Key words: Internet of Things, Machine Learning, Microcontroller