Preface
Author: Umema Sultan
When I first encountered a humanoid robot attempting to walk, I witnessed something remarkable: a machine failing spectacularly, falling repeatedly, yet somehow learning from each collapse. Within hours, chaotic flailing had transformed into tentative steps. Within days, confident strides. I was watching intelligence emerge from experience—not programmed, but discovered.
That moment crystallized the vision behind this textbook.
Why This Book Exists
We stand at an inflection point in robotics. For decades, robots excelled in structured environments—factory floors with precise fixtures, warehouses with predictable layouts, surgical suites with controlled conditions. These machines were marvels of engineering, but they were also brittle, requiring the world to conform to their limitations.
Physical AI changes this equation. By combining modern machine learning with embodied systems, we can build robots that adapt to the world rather than demanding the world adapt to them. Humanoid robots represent the ultimate expression of this vision: machines capable of operating anywhere humans can, using tools designed for human hands, navigating spaces built for human bodies.
Yet the knowledge required to build such systems remains scattered across disciplines and buried in research papers. A student interested in humanoid robotics must piece together insights from control theory, computer vision, reinforcement learning, mechanical engineering, and software architecture. The path from curiosity to competence is unnecessarily difficult.
This textbook aims to clear that path.
My Motivation
I wrote this book because I believe Physical AI will reshape our world in profound ways. Robots that can truly assist humans—in homes, hospitals, disaster zones, and workplaces—will extend human capability and address challenges we cannot solve alone.
But I also wrote it because the field is beautiful. There is elegance in watching a learned policy produce natural walking motion. There is satisfaction in seeing sensor fusion transform noisy data into precise state estimates. There is wonder in a robot that understands spoken commands and translates them into physical action.
I want more people to experience this beauty and contribute to the field. The problems are hard enough without artificial barriers to entry. Every student who struggles to understand ROS conventions, every researcher who reinvents standard techniques, every engineer who makes avoidable architectural mistakes—these represent lost potential.
This textbook is my attempt to lower barriers and accelerate learning.
Philosophy of This Book
Several principles guided the writing:
Concepts over equations. Mathematics is essential for deep understanding, but it can obscure intuition when introduced too early. This book emphasizes conceptual clarity, building understanding that makes later mathematical treatment meaningful rather than mysterious.
Practical grounding. Every concept connects to real implementation. Abstract ideas are anchored in concrete tools: ROS 2, Gazebo, Isaac Sim, actual robot systems. The reader should finish each module ready to build, not just to understand.
Honest complexity. Robotics is genuinely difficult. This book does not pretend otherwise. But it distinguishes essential complexity (the hard problems that define the field) from accidental complexity (the confusing conventions and fragmented tooling that impede progress).
Forward-looking foundations. The field evolves rapidly. Rather than cataloging current tools that may soon be obsolete, this book teaches ways of thinking that remain valuable as technology advances.
Acknowledgments
No textbook emerges from a single mind. I am indebted to the researchers whose papers form the intellectual foundation of this work, to the open-source communities that built the tools we use, and to the students whose questions revealed what needed clearer explanation.
I am grateful to colleagues who reviewed drafts, caught errors, and suggested improvements. Their feedback sharpened arguments and clarified exposition throughout.
Most importantly, I thank the broader robotics community for creating a field worth writing about. The collaborative spirit that shares code, publishes methods, and welcomes newcomers makes this work possible.
An Invitation
This textbook is an invitation to join a community building the future of intelligent machines. The challenges are immense, but so are the opportunities. Every page aims to bring you closer to contributing meaningfully to this endeavor.
The robots are learning to walk. Let us learn to build them together.
Umema Sultan
Next: Introduction →