AMR From Scratch ROS 2, SLAM and Nav2 in Simulation

Build an AMR from an empty file, map a warehouse with SLAM, and navigate it with Nav2 - all in ROS 2 and Gazebo.
What you'll learn

---

Build an autonomous mobile robot from an empty file: URDF, xacro, meshes, wheels, sensors and a working diff-drive plugin in Gazebo.
Map an unseen warehouse with SLAM, then localise inside it by fusing wheel odometry, IMU and laser scans with robot_localization and AMCL.
Configure and tune the full Nav2 stack: costmaps, behaviour trees, planners, controllers and recovery behaviours, on a robot you built yourself.
Derive differential-drive kinematics, odometry and path-planning algorithms on screen, term by term, instead of copying finished equations.
Requirements

---

A computer that can run Ubuntu 24.04. No robot hardware needed
Description

---

This course contains the use of artificial intelligence.
Most robotics courses hand you a finished robot.This one starts with an empty file.
Over thirteen sections, you will build a warehouse-class autonomous mobile robot from scratch usingROS 2, Gazebo, SLAM and Nav2.
You will create the robot Description

---

yourself, including its chassis, wheels, joints, safety scanners, depth camera and IMU, then spawn it inside a realistic Gazebo warehouse and make it navigate autonomously.
What You Will Build
A complete URDF mobile robot
Differential-drive control and kinematics
Laser scanners, depth camera and IMU
Odometry and sensor fusion
SLAM and occupancy-grid mapping
Localization
Nav2 costmaps, planners and controllers
Behavior Trees and recovery behaviors
A complete autonomous warehouse navigation system
Real Code, Real Failures
Every command is actually run on screen.
When something breaks, you watch it fail, inspect the problem and fix it properly. The debugging is intentionally kept in the course because those failures are part of building real robotics systems.
Understand the Mathematics
The mathematics is worked through step by step rather than simply showing finished formulas.
You will understand the foundations behind
Differential-drive kinematics
Odometry integration
Occupancy-grid mapping
Localization
Navigation and path planning
Final Capstone
By the end of the course, all of the pieces come together into one autonomous mobile robot capable of mapping its environment, estimating its position, planning paths, avoiding obstacles and navigating through a warehouse.
Everything runs in simulation.
You need a machine capable of running Ubuntu. You do not need a physical robot.
Who this course is for

---

A computer that can run Ubuntu 24.04. No robot hardware needed
ROS 2 users who can launch someone else's navigation stack but not yet debug their own.
Homepage

---