Course Overview

This course spans 12 weeks across 4 modules. Each module focuses on a critical layer of the Physical AI stack.

Module Map

Module	Topic	Weeks	Key Technologies
1	The Robotic Nervous System	1–3	ROS 2, DDS, Nav2
2	The Digital Twin	4–6	Gazebo, Unity, Sim-to-Real
3	The AI-Robot Brain	7–9	NVIDIA Isaac, RL, Manipulation
4	Vision-Language-Action	10–12	VLA Models, Foundation Models

Module 1: The Robotic Nervous System (Weeks 1–3)

Learning Objectives:

• Understand the ROS 2 architecture and communication patterns
• Build custom ROS 2 nodes, topics, services, and actions
• Implement robot navigation using Nav2

Chapters:

• Chapter 1: ROS 2 Foundations — nodes, topics, publishers, subscribers
• Chapter 2: Advanced ROS 2 — actions, lifecycle nodes, Nav2

Module 2: The Digital Twin (Weeks 4–6)

Learning Objectives:

• Create robot simulations in Gazebo
• Build high-fidelity digital twins in Unity
• Transfer learned behaviors from simulation to physical robots

Chapters:

• Chapter 1: Gazebo Simulation — world building, sensors, physics
• Chapter 2: Unity Digital Twin — rendering, ML-Agents, photorealism
• Chapter 3: Sim-to-Real Transfer — domain randomization, reality gap

Module 3: The AI-Robot Brain (Weeks 7–9)

Learning Objectives:

• Set up NVIDIA Isaac Sim for robot training
• Train manipulation policies using reinforcement learning
• Deploy trained models on physical robot hardware

Chapters:

• Chapter 1: Isaac Sim Fundamentals — setup, environments, task definition
• Chapter 2: Isaac for Manipulation — RL training, grasping, deployment

Module 4: Vision-Language-Action (Weeks 10–12)

Learning Objectives:

• Understand the VLA architecture and how it unifies perception, language, and action
• Explore state-of-the-art VLA models (RT-2, Octo, OpenVLA)
• Fine-tune and deploy VLA models for custom robot tasks

Chapters:

• Chapter 1: VLA Architecture — foundations, multimodal transformers, tokenization
• Chapter 2: VLA Training & Deployment — datasets, fine-tuning, real-world deployment

Assessment

Each module includes embedded code exercises and knowledge checks. The RAG chatbot is available throughout to answer questions and provide hints.

Prerequisites

• Basic Python programming (variables, functions, classes)
• Familiarity with the Linux command line
• Basic understanding of linear algebra and probability (helpful but not required)