The Nav2 loopback simulator is a stand-alone simulator to create a "loopback" for non-physical simulation to replace robot hardware, physics simulators (Gazebo, Bullet, Isaac Sim, etc). It computes the robot's odometry based on the command velocity's output request to create a perfect 'frictionless plane'-style simulation for unit testing, system testing, R&D on higher level systems, and testing behaviors without concerning yourself with localization accuracy or system dynamics.
This was created by Steve Macenski of Open Navigation LLC and donated to Nav2 by the support of our project sponsors. If you rely on Nav2, please consider supporting the project!
It is drop-in replacable with AMR simulators and global localization by providing:
- Map -> Odom transform
- Odom -> Base Link transform,
nav_msgs/Odometry
odometry - Accepts the standard
/initialpose
topic for transporting the robot to another location
Note: This does not provide sensor data, so it is required that the global (and probably local) costmap contain the StaticLayer
to avoid obstacles.
It is convenient to be able to test systems by being able to:
- Arbitrarily transport the robot to any location and accurately navigate without waiting for a particle filter to converge for testing behaviors and reproducing higher-level issues
- Write unit or system tests on areas that are not dependent on low-level controller or localization performance without needing to spin up a compute-heavy process like Gazebo or Isaac Sim to provide odometry and sensor data, such as global planning, autonomy behavior trees, etc
- Perform R&D on various sensitive systems easily without concerning yourself with the errors accumulated with localization performance or imperfect dynamic models to get a proof of concept started
- Simulate N robots simultaneously with a lower compute footprint
- When otherwise highly compute constrained and need to simulate a robotic system
ros2 run nav2_loopback_sim loopback_simulator # As a node, if in simulation
ros2 launch nav2_loopback_sim loopback_simulation.launch.py # As a launch file
ros2 launch nav2_bringup tb3_loopback_simulation.launch.py # Nav2 integrated navigation demo using it
ros2 launch nav2_bringup tb4_loopback_simulation.launch.py # Nav2 integrated navigation demo using it
update_duration
: The duration between updates (default 0.01 -- 100hz)base_frame_id
: The base frame to use (defaultbase_link
)odom_frame_id
: The odom frame to use (defaultodom
)map_frame_id
: The map frame to use (defaultmap
)scan_frame_id
: The can frame to use to publish a scan to keep the collision monitor fed and happy (defaultbase_scan
for TB3,rplidar_link
for TB4)enable_stamped_cmd_vel
: Whether cmd_vel is stamped or unstamped (i.e. Twist or TwistStamped). Defaultfalse
forTwist
.scan_publish_dur
: : The duration between publishing scan (default 0.1s -- 10hz)publish_map_odom_tf
: Whether or not to publish tf frommap_frame_id
toodom_frame_id
(defaulttrue
)publish_clock
: Whether or not to publish simulated clock to/clock
(defaulttrue
)
This node subscribes to:
initialpose
: To set the initial robot pose or relocalization request analog to other localization systemscmd_vel
: Nav2's output twist to get the commanded velocity
This node publishes:
clock
: To publish a simulation clock for all other nodes withuse_sim_time=True
odom
: To publish odometry from twisttf
: To publish map->odom and odom->base_link transformsscan
: To publish a range laser scan sensor based on the static map