Humanoid and Legged Robots: An Introduction
How Optimus, Atlas and robot dogs walk, balance and learn.
Taught by Sristi Maharjan · Robotics Engineer
What you'll learn
- Explain why legged robots exist, what they are good at, and where wheels still win
- Describe how electric actuators, gearboxes and torque sensing produce controlled joint motion
- Compute a simple centre of mass and apply the zero moment point rule for balance
- Distinguish gait types such as walking, trotting and bounding and read a gait diagram
- Outline how IMUs, cameras and state estimation let a robot know where its body is
- Summarise how reinforcement learning and sim-to-real transfer teach humanoids new skills
About this course
Humanoid robots and quadruped robot dogs have moved from research videos into warehouses, factories and pilot programmes, and this beginner course explains how they actually work. It is written for students, engineers from other fields and curious professionals who want a clear, jargon-free understanding of legged robotics before committing to deeper study. You will learn how electric actuators produce joint torque, why balance comes down to the centre of mass and the zero moment point, how gait generators and model predictive control turn footsteps into walking, and how cameras, IMUs and state estimation keep a robot upright on uneven ground. You will also see how embodied AI and reinforcement learning in simulation teach humanoids new skills, and survey the companies building Optimus, Atlas, Figure, Digit and the Unitree platforms. No maths beyond school algebra and a little trigonometry is required; short Python snippets illustrate key ideas.
Course content
6 lessons- 1Why Legs, Why Now: The Case and the LandscapeWhat legged robots are for, the honest trade-offs against wheels, and who is building them in 2026.Preview
- 2Actuators: The Muscles of a Legged RobotElectric motors, gear ratios, torque sensing and why actuator design decides what a robot can do.
- 3Balance: Centre of Mass, Support Polygons and ZMPThe physics of not falling over, from centre of mass to the zero moment point.
- 4Gait: How Robots Walk, Trot and RecoverGait cycles, footstep planning, model predictive control and what happens when a robot is pushed.
- 5Perception and State EstimationHow IMUs, joint encoders, cameras and depth sensors tell a robot where its body and the world are.
- 6Embodied AI: Learning to Walk, Work and What Comes NextReinforcement learning in simulation, sim-to-real transfer, foundation models and your study path.