Include(Menus/MapBuildingStack) RosdocHeader(gmapping) TableOfContents(3)

Package Summary

The gmapping package provides laser-based SLAM (Simultaneous Localization and Mapping). Using slam_gmapping, you can create a 2-D occupancy grid map (like a building floorplan) from laser and pose data collected by a mobile robot.

This is mostly a third party package; the underlying GMapping library is [http://openslam.org/gmapping.html external documented]. Look there for details on many of the parameters listed below.

To use gmapping, you need a mobile robot that provides odometry data and is equipped with a horizontally-mounted, fixed, laser range-finder. gmapping will attempt to transform each incoming scan into the odometry frame; make sure that all necessary transforms are available. Normally there are two transforms involved:

• base_link -> odom : provided by the odometry system (e.g., the driver for the mobile base)

• base_link -> base_laser : a fixed value, broadcast periodically by mechanism_control, or a tf/transform_sender

Example

ROS API

Subscribed topics

• scan (sensor_msgs/LaserScan) : Laser scans to create the map from

Advertised services

• dynamic_map (nav_msgs/GetMap) : Call this service to get the map data

Parameters used

• ~inverted_laser (default: false) : Is the laser right side up (scans are ordered CCW), or upside down (scans are ordered CW)?
• ~throttle_scans (default: 1) : Process 1 out of every this many scans (set it to a higher number to skip more scans)
• ~base_frame (default: "base_link") : The frame attached to the mobile base.
• ~map_frame (default: "map") : The frame attached to the map.
• ~odom_frame (default: "odom") : The frame attached to the odometry system.
• ~map_update_interval_ (default: 5.0) : How long (in seconds) between updates to the map. Lowering this number updates the occupancy grid more often, at the expense of greater computational load.
• ~maxUrange (default: 80.0) : The maximum usable range of the laser. A beam is cropped to this value.
• ~sigma (default: 0.05) : The sigma used by the greedy endpoint matching
• ~kernelSize (default: 1) : The kernel in which to look for a correspondence
• ~lstep (default: 0.05) : The optimization step in translation
• ~astep (default: 0.05) : The optimization step in rotation
• ~iterations (default: 5) : The number of iterations of the scanmatcher
• ~lsigma (default: 0.075) : The sigma of a beam used for likelihood computation
• ~ogain (default: 3.0) : Obstacle gain (?)
• ~lskip (default: 0) : Number of beams to skip in each scan.
• ~srr (default: 0.1) : Odometry error in translation as a function of translation (rho/rho)
• ~srt (default: 0.2) : Odometry error in translation as a function of rotation (rho/theta)
• ~str (default: 0.1) : Odometry error in rotation as a function of translation (theta/rho)
• ~stt (default: 0.2) : Odometry error in rotation as a function of rotation (theta/theta)
• ~linearUpdate (default: 1.0) : Process a scan each time the robot translates this far
• ~angularUpdate (default: 0.5) : Process a scan each time the robot rotates this far
• ~resampleThreshold (default: 0.5) : The neff based resampling threshold (?)
• ~particles (default: 30) : Number of particles in the filter
• ~xmin (default: -100.0) : Initial map size
• ~ymin (default: -100.0) : Initial map size
• ~xmax (default: 100.0) : Initial map size
• ~ymax (default: 100.0) : Initial map size
• ~delta (default: 0.05) : Processing parameters (resolution of the map)
• ~llsamplerange (default: 0.01) : Translational sampling range for the likelihood
• ~llsamplestep (default: 0.01) : Translational sampling range for the likelihood
• ~lasamplerange (default: 0.005) : Angular sampling range for the likelihood
• ~lasamplestep (default: 0.005) : Angular sampling step for the likelihood

tf API

Required transforms

• (the frame attached to incoming scans) -> base_link

• base_link -> odom

Provided transforms

• map - > odom

[wiki:/Reviews Review Process]