A node that maintains a local obstacle map (point cloud, voxels or occupancy grid) from recent sensor readings within a configurable time window.

Published Topics

local_map_pointcloud (sensor_msgs/PointCloud)
  • The local map as a point cloud. The frame_id is <frameid_robot> (default: odom).


~frameid_reference (string, default: "odom")
  • TF frame name for the reference frame
~frameid_robot (string, default: "base_link")
  • TF frame name for the robot
~source_topics_2dscan (string, default: "scan,laser1") ~topic_local_map_pointcloud (string, default: "local_map_pointcloud")
  • Name of the topic where to publish the local map
~time_window (double, default: 0.2)
  • For how long past observations will be stored in the local map (seconds)
~publish_period (double, default: 0.1)
  • How often to publish the local map (seconds)
~show_gui (bool, default: true)
  • Show an independent window with a 3D view of the local obstacles map

Required tf Transforms

<value of ~frameid_robot> (typ: base_link)<each sensor frameid> (typ: scan, etc)
  • local pose of each sensor on the robot
<value of ~frameid_reference> (typ: odom)<value of ~frameid_robot> (typ: base_link)
  • pose of the robot on the reference frame


This demo requires mvsim. It simulates one robot with 2 laser scanners and builds and publish the local obstacle maps to /local_map_pointcloud:

roslaunch mrpt_local_obstacles demo_with_mvsim.launch


Wiki: mrpt_local_obstacles (last edited 2015-04-28 23:33:45 by Jose Luis Blanco)