Lessons
Each lesson is a series of short programming steps, most of them driving a live physics simulation. Your progress is saved in this browser.
Foundations
Free refreshers on the maths and physics the lessons use, each step on the same simulated arm. Take them first, or jump to a step when a lesson needs it: they open in any order.
Courses
Beginner, about 35 minutes
Python Basics
Never programmed before? Learn the Python the robot lessons use, one short program at a time.
Beginner, about 50 minutes
C++ Basics
Never programmed before, or new to C++? Learn the C++ the robot lessons use, one short program at a time.
Beginner, about 1 hour
Pick & Place
Program a robot arm to clear a table of coloured balls, the way real manipulation code is written.
Intermediate, about 1 hour
Jacobians & Velocity Control
Control the gripper's velocity instead of its position: Jacobians, resolved-rate control and numerical IK.
Intermediate, about 1 hour
Trajectory Generation
Plan how the arm moves over time: polynomials, velocity limits and smooth paths through waypoints.
Intermediate, about 75 minutes
Motion Planning
Get around obstacles: configuration space, collision checking, RRT and roadmaps.
Intermediate, about 75 minutes
Seeing with a Camera
Replace perfect knowledge with a camera: project, segment, back-project, then pick what you see.
Intermediate, about 1.7 hours
Depth & Point Clouds
See the table in 3-D: turn depth into point clouds, find the table with RANSAC, then pick objects by their shape.
Intermediate, about 75 minutes
Noisy Sensors & Kalman Filters
Real sensors lie a little. Average, fuse and track, then catch a ball that won't sit still.
Intermediate, about 75 minutes
Mobile Robot Navigation
Odometry, occupancy grids and path planning for a wheeled robot.
Advanced, about 1.5 hours
Localisation & SLAM
Where am I? Filters, scan matching and loop closure for a robot that can't trust its wheels.
Intermediate, about 75 minutes
Dynamics & Torque Control
Drive the arm with raw motor torques: equations of motion, gravity compensation, PID, computed torque and compliance.
Advanced, about 75 minutes
Balancing & Optimal Control
Keep an inverted pendulum upright: linear models, eigenvalues, pole placement, LQR and MPC on a cart-pole.
Advanced, about 1.7 hours
Robot Learning
Let the robot learn its own controller: value iteration, Q-learning, imitation, policy gradients and model-based RL.
Intermediate, about 1.8 hours
ROS 2 for Robot Arms
Write real ROS 2 nodes for a robot arm, from your first topic to trajectory actions, with nothing to install.
Want one algorithm at a time? Try the Problems: a short exercise for each core robotics algorithm.
Want something you're not seeing?
Tell us which lesson to build next.
Just getting started? Begin with Python Basics or C++ Basics, or go straight to Pick & Place if you already know Python or C++.