New in C Turtle

melodic noetic   Show EOL distros: 

Documentation

Part of the documentation is available in the previously maintained karto package.

Nodes

slam_karto

The slam_karto node takes in sensor_msgs/LaserScan messages and builds a map (nav_msgs/OccupancyGrid). The map can be retrieved via a ROS topic or service.

Subscribed Topics

tf (tf/tfMessage)

• Transforms necessary to relate frames for laser, base, and odometry (see below)

scan (sensor_msgs/LaserScan)

• Laser scans to create the map from

Published Topics

map_metadata (nav_msgs/MapMetaData)

• Get the metadata of the map data (resolution, width, height, ...)

map (nav_msgs/OccupancyGrid)

• Get the map data from this topic, which is latched, and updated periodically

visualization_marker_array (visualization_msgs/MarkerArray)

• Get the pose graph from this topic, which is updated periodically

Services

dynamic_map (nav_msgs/GetMap)

• Call this service to get the map data

Parameters

~odom_frame (string, default: "odom")

• The frame attached to the odometry system.

~map_frame (string, default: "map")

• The frame attached to the map.

~base_frame (string, default: "base_link")

• The frame attached to the mobile base.

~throttle_scans (int, default: 1)

• Process 1 out of every this many scans (set it to a higher number to skip more scans)

~map_update_interval (float, 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.

~resolution (float, default: 0.05)

• Resolution of the map (in metres per occupancy grid block)

~delta (float, default: 0.05)

• Resolution of the map (in metres per occupancy grid block). Same as resolution. Defined for compatibility with the parameter names of gmapping.

~transform_publish_period (float, default: 0.05)

• How long (in seconds) between transform publications. To disable broadcasting transforms, set to 0.

~use_scan_matching (bool, default: true)

• When set to true, the mapper will use a scan matching algorithm. In most real-world situations this should be set to true so that the mapper algorithm can correct for noise and errors in odometry and scan data. In some simulator environments where the simulated scan and odometry data are very accurate, the scan matching algorithm can produce worse results. In those cases set this to false to improve results.

~use_scan_barycenter (bool, default: true)

• Use the barycenter of scan endpoints to define distances between scans.

~minimum_travel_distance (double, default: 0.2)

• Sets the minimum travel between scans. If a new scan's position is more than minimumTravelDistance from the previous scan, the mapper will use the data from the new scan. Otherwise, it will discard the new scan if it also does not meet the minimum change in heading requirement. For performance reasons, generally it is a good idea to only process scans if the robot has moved a reasonable amount.

~minimum_travel_heading (double, default: deg2rad(10))

• Sets the minimum heading change between scans. If a new scan's heading is more than minimum_travel_heading from the previous scan, the mapper will use the data from the new scan. Otherwise, it will discard the new scan if it also does not meet the minimum travel distance requirement. For performance reasons, generally it is a good idea to only process scans if the robot has moved a reasonable amount.

~scan_buffer_size (int, default: 70)

• Sets the length of the scan chain stored for scan matching. scan_buffer_size should be set to approximately scan_buffer_maximum_scan_distance / minimum_travel_distance. The idea is to get an area approximately 20 meters long for scan matching. For example, if we add scans every minimum_travel_distance == 0.3 meters, then scan_buffer_size should be 20 / 0.3 = 67.

~scan_buffer_maximum_scan_distance (double, default: 20.)

• Sets the maximum distance between the first and last scans in the scan chain stored for matching.

~link_match_minimum_response_fine (double, default: 0.8)

• Scans are linked only if the correlation response value is greater than this value.

~link_scan_maximum_distance (double, default: 10.)

• Sets the maximum distance between linked scans. Scans that are farther apart will not be linked regardless of the correlation response value.

~loop_search_maximum_distance (double, default: 4.)

• Scans less than this distance from the current position will be considered for a match in loop closure.

~do_loop_closing (bool, default: true)

• Enable/disable loop closure.

~loop_match_minimum_chain_size (int, default: 10)

• When the loop closure detection finds a candidate it must be part of a large set of linked scans. If the chain of scans is less than this value, we do not attempt to close the loop.

~loop_match_maximum_variance_coarse (double, default: sqrt(0.4))

• The co-variance values for a possible loop closure have to be less than this value to consider a viable solution. This applies to the coarse search.

~loop_match_minimum_response_coarse (double, default: 0.8)

• If response is larger than this, then initiate loop closure search at the coarse resolution.

~loop_match_minimum_response_fine (double, default: 0.8)

• If response is larger than this, then initiate loop closure search at the fine resolution.

~correlation_search_space_dimension (double, default: 0.3)

• Sets the size of the search grid used by the matcher. The search grid will have the size correlation_search_space_dimension x correlation_search_space_dimension.

~correlation_search_space_resolution (double, default: 0.01)

• Sets the resolution (size of a grid cell) of the correlation grid.

~correlation_search_space_smear_deviation (double, default: 0.03)

• The point readings are smeared by this value in X and Y to create a smoother response.

~loop_search_space_dimension (double, default: 8.0)

• The size of the search grid used by the matcher.

~loop_search_space_resolution (double, default: 0.05)

• The resolution (size of a grid cell) of the correlation grid.

~loop_search_space_smear_deviation (double, default: 0.03)

• The point readings are smeared by this value in X and Y to create a smoother response.

~distance_variance_penalty (double, default: sqrt(0.3))

• Variance of penalty for deviating from odometry when scan-matching. The penalty is a multiplier (less than 1.0) is a function of the delta of the scan position being tested and the odometric pose.

~angle_variance_penalty (double, default: sqrt(deg2rad(20)))

• See distance_variance_penalty.

~fine_search_angle_offset (double, default: deg2rad(0.2))

• The range of angles to search during a fine search.

~coarse_search_angle_offset (double, default: deg2rad(20.0))

• The range of angles to search during a coarse search.

~coarse_angle_resolution (double, default: deg2rad(2.0))

• Resolution of angles to search during a coarse search.

~minimum_angle_penalty (double, default: 0.9)

• Minimum value of the angle penalty multiplier so scores do not become too small.

~minimum_distance_penalty (double, default: 0.5)

• Minimum value of the distance penalty multiplier so scores do not become too small.

~use_response_expansion (bool, default: false)

• Whether to increase the search space if no good matches are initially found

Required tf Transforms

<the frame_id of the incoming scans> → base_link

• usually a fixed value, broadcast periodically by a robot_state_publisher, or a tf static_transform_publisher.

base_link → odom

• usually provided by the odometry system (e.g., the driver for the mobile base)

Provided tf Transforms

map → odom

• the current estimate of the robot's pose within the map frame

Report a Bug

Use github to report a bug or propose an enhancement.