CLA — ME 396P Nav2 Demo

This demo runs the TurtleBot3 Navigation2 stack in simulation using Docker and tmux automation scripts.
The workflow is designed to run entirely within one container session — you enter the container once, perform SLAM to create a map, then run Navigation2 using that saved map. You can then run Nav2 with a custom python script for and automated survey demo.

---

Pre-Requisites

Docker Setup

Install Docker and Docker Compose on your host system:
https://www.docker.com/get-started/

Verify installation:

docker --version
docker compose version

Environment Requirements

• Host: X11 display enabled (default on most desktops)
• GPU acceleration is optional; CPU simulation is sufficient for this demo.

---

Getting Started

GUI Permission (on host)

Allow Docker containers to display graphical windows (Gazebo and RViz):

xhost +local:root

Build the Docker Image (on host)

From the project root directory (where your Dockerfile and docker-compose.yml are located):

docker compose up --build -d

This builds the container image and installs:

• ROS 2 Humble
• TurtleBot3 Simulation Packages
• Navigation2 (Nav2)
• Cartographer SLAM
• tmux + automation scripts

Enter the Container

Once the container is running:

docker exec -it nav2-demo bash

You should now see something like:

root@<container_id>:/workspace/src#

You are now inside the container’s shell environment, starting in the workspace/src directory.

All subsequent steps take place inside this container.

To leave the container at any time, simply type:

exit

This will return you to your host machine’s terminal.

---

Stage 1 — SLAM (Cartographer)

Launch SLAM

Inside the container:

bash /root/scripts/slam_tmux.sh

This starts a tmux session that automatically launches:

• Gazebo simulation
• Cartographer (SLAM) + RViz
• Teleop control
• Map saver utility

Tab Layout (Shown at Bottom of Terminal)

Tab	Name	Description
[0]	guide	Instructions + quit control (q)
[1]	gazebo	Gazebo launch
[2]	slam	Cartographer (SLAM) + RViz launch
[3]	teleop	Keyboard control (must be active tab to drive)
[4]	map_saver	Press Enter anytime to save map

Navigating Between Tabs

Click the tab name at bottom of the terminal OR use the Alt + # shortcuts:

• Alt + 0 → guide
• Alt + 1 → gazebo
• Alt + 2 → slam
• Alt + 3 → teleop
• Alt + 4 → map_saver

Operating SLAM

1. Start in [0] guide to review instructions.
2. Wait for Gazebo and RViz to pop-up.
3. Switch to [3] teleop.
4. Before driving:

   • Be able to see [3] teleop terminal, RViz, and Gazebo.

   • Align the starting RViz map with Gazebo to the best of your ability.

   • Suggested Single Monitor Layout:

     

5. How to drive:
   • Be selected into [3] teleop terminal.
   • Use W/A/S/D/X keys to control the robots linear and angular velocities.
   • *Controlling velocities to navigate can feel weird at first, remember you can always stop with S.
6. As you move the robot:
   • Gazebo shows the simulated robot moving in the environment.
   • RViz Displays the occupancy grid map updating in real time
     • Walls and obstacles appear as black lines.
     • Unexplored regions are gray.
     • Free space is white.
7. When the occupancy map appears complete:

   • Example of a completed map:

     

   • Switch to [4] map_saver.

   • Press Enter to save the map to:

     /root/data/my_map.yaml
     

8. Once finished, return to [0] guide and enter q to quit the session.
   This stops all processes and returns you to the container shell.

---

Stage 2 — Navigation (Using Saved Map)

After completing SLAM and saving the map, you can launch the Navigation2 stack.

Launch Navigation2

Inside the same container session:

bash /root/scripts/nav_tmux.sh

This starts a tmux session that launches:

• Gazebo simulation
• Nav2 + RViz

Tab Layout (Shown at Bottom of Terminal)

Tab	Name	Description
[0]	guide	Instructions + quit control (q)
[1]	gazebo	Gazebo launch
[2]	nav2	Nav2 + RViz launch

Navigating Between Tabs

Use the Alt + number shortcuts:

• Alt + 0 → guide
• Alt + 1 → gazebo
• Alt + 2 → nav2

Operating Nav2

1. Start in [0] guide to review instructions.
2. Wait for Gazebo and RViz to pop-up.

   • RViz should now display the saved map from the SLAM stage.

   • View RViz and Gazebo side by side.

   • Suggested Single Monitor Layout with 2D Pose Estimate:

     

3. In RViz use the 2D Pose Estimate tool to set the robot's initial pose:
   • Click the 2D Pose Estimate tool in the top tool bar.
   • Click and drag on the map to indicate the robot’s approximate position and heading (can reference Gazebo for rough initial pose).
4. Once localization is stable, use the Nav2 Goal tool to send navigation goals:
   • Click the Nav2 Goal button in the top toolbar.
   • Click and drag on the desired destination on the map to define the goal position and orientation.
   • The robot will plan a path (displayed as a colored line) and begin moving toward that goal.
   • This can be repeated to send additional goals.
5. You can monitor progress with Gazebo and RViz.
   • Gazebo shows the robot "physically" moving around the simulated world.
   • RVIZ shows the planned path, current postion, local costmap, and goal status updates in real time.
6. When finished, return to [0] guide terminal and enter q to quit the session.
   This stops all processes and returns you to the container shell.

Stage 3 — Automated Navigation Survey

After manually tasking your in Nav2, you can now see how you can automate that process with Python!

Launch Nav2 + Survey

Inside the same container session:

bash /root/scripts/survey_tmux.sh

This starts a tmux session that launches:

• Gazebo simulation
• Nav2 + RViz
• Python survey script

Tab Layout (Shown at Bottom of Terminal)

Tab	Name	Description
[0]	guide	Instructions + quit control (q)
[1]	gazebo	Gazebo launch
[2]	nav2	Nav2 + RViz launch
[3]	survey	Run survey script

Navigating Between Tabs

Use the Alt + number shortcuts:

• Alt + 0 → guide
• Alt + 1 → gazebo
• Alt + 2 → nav2
• Alt + 3 → survey

Operating Nav2 Survey

1. Start in [0] guide to review instructions.
2. Wait for Gazebo and RViz to pop-up.
   • RViz should now display the saved map from the SLAM stage.
   • View RViz and Gazebo side by side (like in Stage 2)
   • If Gazebo is hung up, just wait, since we are now using a larger map, can take longer to boot up.
3. You can monitor [3] survey terminal for tasking updates, however the survey should start automatically for you to watch!
   • Gazebo shows the robot "physically" moving around the simulated world.
   • RVIZ shows the planned path, current postion, local costmap, and goal status updates in real time.
4. When finished, return to [0] guide terminal and enter q to quit the session.
   This stops all processes and returns you to the container shell.
5. To exit the container:

   exit
   

---

File Overview

File	Purpose
Dockerfile	Builds ROS 2 Humble + TurtleBot3 + Nav2 + Cartographer + tmux
docker-compose.yml	Defines container runtime and GUI configuration
.tmux.conf	Defines Alt + number keybindings and status bar layout
slam_tmux.sh	Automated SLAM (Gazebo + Cartographer + Teleop + Map Saver)
nav_tmux.sh	Automated Navigation2 (Gazebo + Nav2)
survey_tmux.sh	Automated Survey (Gazebo + Nav2 + python script)
nav2_test.py	Python script that sends initial and goal poses

---

Summary

This workflow keeps all operations within a single container session.
You build once, enter once, and use tmux automation to transition seamlessly between SLAM and Navigation without managing multiple terminals.
Mapping, saving, localization, navigation, and a automatic survey demo are all handled through structured tmux windows with simple keyboard shortcuts.