LaMa ROS - Alternative Localization and Mapping for ROS.

https://github.com/iris-ua/iris_lama_ros

Overview

ROS integration of LaMa, a Localization and Mapping package from the Intelligent Robotics and Systems (IRIS) Laboratory, University of Aveiro. It provides 2D Localization and SLAM. It works great on a TurtleBot2 with a Raspberry Pi 3 Model B+ and an Hokuyo (Rapid URG).

Build

To build LaMa ROS, clone it from GitHub and use catkin to build.

mkdir src
cd src
git clone https://github.com/iris-ua/iris_lama
git clone https://github.com/iris-ua/iris_lama_ros
cd ..
catkin config --extend /opt/ros/melodic
catkin build

The build was tested in Ubuntu 18.04 with ROS melodic. It will not build with catkin_make or catkin_make_isolated.

SLAM nodes

To create a map using Online SLAM execute

rosrun iris_lama_ros slam2d_ros scan:=base_scan

and to create a map using Particle Filter SLAM execute

rosrun iris_lama_ros pf_slam2d_ros scan:=base_scan

Both nodes will publish to expected topics such as /map and /tf.

Parameters

~global_frame_id: The frame attached to the map (default: "map").
~odom_frame_id: The frame attached to the odometry system (default: "odometry").
~base_frame_id: The frame attached to the mobile base (default: "base_link")
~d_thresh: Traveled distance to accumulate before updating (default: 0.5 meters).
~a_thresh: Angular motion to accumulate before updating (default: 0.25 rads).
~l2_max: Maximum distance to use in the dynamic Euclidean distance map (default: 0.5 meters).
~use_compression: Should the maps be compressed (default: false).
~compression_algorithm: Compression algorithm to use, lz4 or zstd (default: "lz4").

Particle Filter SLAM only:

~particles: Number of particles to use (default: 30).
~threads: Number of working threads, -1 means disabled and 0 will expand to the available number of cores (default: -1).
~lgain: Gain value for smoothing the particles likelihood (default: 3.0).

Localization node

This node requires the existence of the /static_map service to load the map. To run the localization just execute

rosrun iris_lama_ros loc2d_ros scan:=base_scan