SLAM: building a map from scratch with slam_toolbox

This is a valid v1.0 placeholder page for the later curriculum arc. Full interactive lab treatment ships after Week 1 dogfooding.

LECTURE & READING

Glossary primer (10 min)

— Simultaneous and . Build a map of an unknown while tracking your pose in it.
_toolbox — ROS 2's go-to 2D LiDAR-based . Maintained by SteveMacenski. Replaced gmapping.
Pose graph — 's internal representation: nodes are poses, edges are scan-match constraints.
Loop closure — Detecting "I've been here before" and correcting the accumulated drift.
Occupancy grid — Output map. Cells are free / occupied / unknown.
Localization-only mode — Use a saved map; track pose without modifying the map.
Lifelong — Continuous that adds new observations to a saved map.

Real-world analogy

is sketching a building's floor plan while walking through it blindfolded — you remember "8 paces forward, 5 right, hit a wall" and combine that with "this corner looks like one I've seen before" to keep the map consistent.

Hour 1 — Stachniss SLAM primer

This is the visual primer for . Watch ~30 min total:

Cyrill Stachniss sequence:
, 5 Minutes with Cyrill
Video
, 5 Minutes with Cyrill
Open source
Graph-based using Pose Graphs
Video
Graph-based using Pose Graphs
Open source
Bag of Visual Words, 5 Minutes with Cyrill
Video
Bag of Visual Words, 5 Minutes with Cyrill
Open source
Pick the videos titled "What is ?", "Pose Graph ", "Loop Closure".

Stachniss is the world's best at making intuitive. Don't skip.

Hour 2 — slam_toolbox docs

README and design overview: https://github.com/SteveMacenski/slam_toolbox
Articulated Robotics' video: https://articulatedrobotics.xyz/tutorials/mobile-robot/concept/concept-slam/

LAB

Hour 3 — Lab: build a warehouse map, save it, re-localize against it (75 min)

What you're building. Run _toolbox in online async mode while teleoping the TurtleBot4 around the warehouse. Save the resulting map. Then re-launch in localization-only mode and verify the saved map can be reused with AMCL.

What success looks like at the end. You have: 1. A .pgm and .yaml map pair saved at ~/maps/my_warehouse.{pgm,yaml}. 2. figures/day12_warehouse_map.png showing the map opened in any image viewer. 3. The map covers ≥ 80% of the warehouse (visually; no large unmapped regions). 4. After re-launch in mode, AMCL converges using the saved map.

Step 1 — Launch TB4 sim + slam_toolbox (10 min)

Pane 1:

Full source continues in the committed curriculum files. The v1.0 page exposes the day flow and lab surface without inventing content.

Completion controls unlock when this day graduates from placeholder to full lab.

LECTURE & READING

Glossary primer (10 min)

Real-world analogy

Hour 1 — Stachniss SLAM primer

SLAMNavigation & LocomotionSLAMSimultaneous Localization and Mapping., 5 Minutes with Cyrill

Graph-based SLAMNavigation & LocomotionSLAMSimultaneous Localization and Mapping. using Pose Graphs

Bag of Visual Words, 5 Minutes with Cyrill

Hour 2 — slam_toolbox docs

LAB

Hour 3 — Lab: build a warehouse map, save it, re-localize against it (75 min)

Step 1 — Launch TB4 sim + slam_toolbox (10 min)

, 5 Minutes with Cyrill

Graph-based using Pose Graphs