Skip to main content

Listed

, home
FREN

Nav2

Definition

The ROS 2 navigation stack, which succeeds the ROS Navigation Stack. It provides perception of the environment, path planning, computation of motor commands, obstacle avoidance and localization, for mobile robots of varied kinematics, size and equipment. Its logic is described by behavior trees.

Checked on against the references below. · Written with AI · checked against the sources Source: Nav2 documentation.

Concrete examples

On Stretch 4, Hello Robot’s repository gives Nav2 the job of “planning and collision avoidance”. GLIM provides lidar-inertial odometry, and RTAB-Map handles the map and localization. For Unitree Go2, a community repository lists obstacle avoidance by Nav2 using sensor data (“Obstacle avoidance using Nav2 and sensor data”). Another community project, for Unitree G1, includes “Nav2 navigation” in a ROS 2 Jazzy stack developed first in simulation on unitree_mujoco.

What it changes for you

Nav2 support depends on the ROS 2 distribution: Lyrical Luth and Jazzy Jalisco are at “Active Support”, Humble Hawksbill and Kilted Kaiju are “Maintained”, and Iron Irwini and Galactic Geochelone are “End of Life”. On the setup side, the Stretch 4 repository requires Ubuntu 24.04 with ROS 2 Jazzy and builds GTSAM, GLIM and a patched RTAB-Map from source.

Not to be confused with

In grove-g1, in simulation, SLAM Toolbox builds the map, then AMCL, a Nav2 package, localizes the robot on the saved map before Nav2 drives it to the goal. On Stretch 4, where RTAB-Map provides the map and localization, goals are cancelled until RTAB-Map is running and a 2D pose estimate has been set.

Additional material checked on .

Related terms

References

Back to the glossary