AI for Robotics - aegean.ai

What this course is all about

This course introduces students to the foundational concepts of AI as they apply to the complex world of robotics, emphasizing how AI algorithms enable robots to interpret sensor data, navigate dynamic environments, and interact intelligently with both objects and humans. The course explores key AI methods, including computer vision, natural language processing, path planning and reinforcement learning that give robots the ability to analyze, reason, and respond to the world around them. Students will engage with hands-on projects with a focus on real-world applications in areas like autonomous vehicles and industrial automation. By the end of the course, students will have gained a deep understanding of how AI enables autonomous robots to perform complex tasks, being instructed with natural language making them capable of working alongside humans in various settings.

Topics Covered

Kinematics

Configuration space and motion representations for robot control.

Global Planning

Search algorithms including forward search and A* for path planning.

Local Planning

Motion planning, behavioral planning, and prediction for dynamic environments.

VLA Agents

Vision-Language-Action models, imitation learning, and sim2real transfer.

Simulation

Simulation environments for training embodied AI agents.

Video Lectures

9 modules · 43 lessons · 10+ hours of content

Subscribe to our YouTube channel and explore the complete curriculum below.

1. Course Introduction

The Robotic AI Agent

A practical map for navigating robotic AI systems.

Development Environment Setup

Setting up the ROS2 Docker-based development environment.

Mathematical Prerequisites

What you need to know before diving into the course material.

2. Statistical Learning Theory

The Learning Problem

The Vapnik block diagram.

Linear Regression

Extracting non-linear patterns with linear models.

Gradient Descent

Optimizing complicated functions with iterative methods.

Entropy

Information theory principles.

Maximum Likelihood Estimation

The workhorse of statistical modeling.

Binary Classification

Binary classification and Logistic Regression.

3. Neural Networks

Feature Extraction

Using a simple network to understand how features are extracted.

Multiclass Classifier

A simple multiclass classifier example.

Backpropagation

How to calculate gradients in a neural network.

Regularization

How to regulate the complexity of a neural network.

4. Convolutional Neural Networks

How We Understand Scenes

Human Perception and Imaging.

Convolution and Correlation

A linear operation for extracting spatial features.

CNN Architectures

Looking inside a CNN layer and understanding architectural patterns.

Image Classification

Image classification with data augmentation.

What CNNs Learn

Visualizing the features learned by CNNs.

ResNets

Residual Networks and skip connections.

5. Object Detection

Introduction to Object Detection

Object detection in a physical security application.

Computer Vision Datasets

What types of annotations are used in computer vision?

Region-based Object Detectors

R-CNN, Fast R-CNN, Faster R-CNN.

6. Recursive State Estimation

Introduction to State Estimation with HMM

Introducing Hidden Markov Models.

The Bayes Filter

Implementing the Bayes filter algorithm.

Discrete Bayes Filter Example

Discrete Bayes localization notebook.

Continuous State Space and Kalman Filter

Localizing a drone under Gaussian assumptions.

A Kalman Filter Example

Kalman localization notebook.

7. Global Planning

Introduction to Planning

Typical planning problems and PDDL.

Planning Domain Definition Language

The constructs of PDDL.

Forward Search Algorithms

Finding global planning solutions with forward search.

The A* Algorithm

Using heuristics to guide forward search.

8. Multimodal Reasoning and Transformers

Introduction to Transformers

The transformer architecture and the simple attention mechanism.

The Learnable Attention Mechanism

Implementing the scaled dot-product self attention mechanism.

Multi-Head Self Attention

Using multiple attention heads to capture different aspects of input sequences.

9. Markov Decision Processes

Introduction to MDPs - Part 1

Defining Markov Decision Processes.

Introduction to MDPs - Part 2