Download as pdf or txt
Download as pdf or txt
You are on page 1of 1

AI & Machine Learning for Embedded Systems (EDGE AI)

(Courseware designed in collaboration with STMicroelectronics)

Introduction to AI & Machine Learning for Embedded Systems


Embedded Systems fundamentals, applications, building blocks &design approach, introduction to AI
& Machine Learning, Traditional vs. Embedded Machine Learning- benefits & design challenges.

STM32 Architecture and Peripherals


Introduction to ARM Cortex M4 Architecture, Functional Block, STM32 Peripherals- Clock System,
GPIO, Timers & Interrupts,
Serial Communication- UART frame format, theory of operation, baud rate configuration, polling &
interrupt method
Analog to Digital Converter 12 bit ADC configuration- theory of operation, ADC polling mode vs.
interrupt mode, DMA with ADC, single channel &multi-channel
I2C communication protocol- theory of operation, frame format, master slave protocol, multi slave
protocol

Sensor Signal Processing & Data Structure for Machine Learning


Introduction to Audio Sensing & Inertial Motion sensors. Understanding sampling rate, aliasing, and
resolution. Understanding techniques to acquire sensor signals. Features extraction for sensor signal
classification: Fast Fourier Transform, average filter, FIR filter, MFCC.
Designing a Data Logger, converting real time sensor data in standard data format(.csv), modelling &
plotting sensor data using matplotlib, processing sensor data- cleaning sensor data a) missing values
b) constant samples, c) random values

Machine Learning Models for Embedded Systems


Understanding Supervised & Unsupervised Machine Learning, Training& Inference in ML,
comparison- ML & non-ML project. Concepts of Over-training & Generalization
Anomaly Detection- definition & types, applications, fundamental statistics used for anomaly
detection: mean, standard deviation, variance.
Classification problems in machine learning- common algorithms, decision trees & support vector
machine. Introduction to Neural Networks, gradient descent, bias, confusion matrix
Regression techniques:
TensorFlow lite framework for microcontrollers, convolutional neural network, classification model
development-training & inference

Case Studies on Machine Learning for Embedded Systems


a. Understanding predictive maintenance applications & machine learning approach in industrial
ecosystem.
b. Understanding the applications of machine learning for efficient health monitoring of
patients.
c. Applications of machine learning for a smart home: predicting energy consumption of home
appliances and heating/cooling requirements of a building.

DigiToad Technologies
AB square #617, 4th Floor, 5th main
OMBR layout Bangalore 560043
www.digitoadtech.com

You might also like