This repository provides a configuration for the navigation of the AgileX Scout robot with NAV2 and a simulation environment for the robot in Ignition Gazebo v6 (Fortress). The robot is a skid-steering mobile robot with a 3D lidar sensor mounted on top. The navigation stack launcher is based on the package nav2_bringup from the Navigation2 stack. The simulation environment is based on the AWS RoboMaker Small Warehouse World, which is a Gazebo world well suited for testing robot navigation in both cluttered and open spaces.

Author: Simone Giampà

Project realized at Politecnico di Milano, Artificial Intelligence and Robotics Laboratory (AIRLAB)

Project part of my Master's Thesis project for the Master's Degree in Computer Science and Engineering

Academic Year: 2023/2024

This repository contains the configuration files and the launch files for performing autonomous navigation with the AgileX Scout robot using Nav2. It also contains the configuration files for the simulation of the robot in Ignition Gazebo v6 (Fortress).

All parameters used for the Navigation stack are collected in several YAML configuration files, inside the scout_nav2/params folder. They have been carefully tuned to work as well as possible with the AgileX Scout robot and the simulation environment. Despite this, it is still recommended to consider the specific characteristics of the environment in which the robot will operate and to keep an eye on the robot during operation for safety reasons.

The following dependencies are required to run the simulation and the navigation stack:

• navigation2
• pointcloud-to-laserscan
• slam-toolbox
• ros-gzfortress
• ignition-gazebo6
• spatio-temporal-voxel-layer

The dependencies can be installed using the following command:

$ sudo apt install ros-<$ros-distro>-<dependency-name>

The compilation of the packages is required to run the simulation and the navigation stack. The packages can be compiled using the following command:

$ colcon build --DCMAKE_BUILD_TYPE=Release

Be aware that the compilation of the spatio-temporal-voxel-layer package can take a long time and consume a lot of memory, due to its dependecy on the openvdb library. If the compilation takes up too much memory, it is possible to limit the number of cores used by the compiler with the following command:

$ MAKEFLAGS="-j4" colcon build --cmake-args -DCMAKE_BUILD_TYPE=Release

Where the number 4 is the number of cores used by the compiler.

The robot can be tested in both simulated and real environments. The following steps are required to launch the robot in both environments. Several parameters are provided in the launch files to customize the robot's behavior and the navigation stack.

The parameters for the navigation stack are collected in several YAML configuration files, inside scout_nav2/params folder. The available configuration files are:

• scout_amcl.yaml: navigation parameters using AMCL for localization and SLAM toolbox for map creation, with real Scout robot
• scout_slam_localization.yaml: navigation parameters using SLAM toolbox for localization only, with real Scout robot
• sim_lidar2d_amcl.yaml: navigation parameters using AMCL for localization and SLAM toolbox for map creation, with simulated Scout robot and 2D lidar
• sim_lidar3d_amcl.yaml: navigation parameters using AMCL for localization and SLAM toolbox for map creation, with simulated Scout robot and 3D lidar
• sim_slam_localization.yaml: navigation parameters using SLAM toolbox for localization only, with simulated Scout robot and 3D lidar

The parameters used for the navigation are automatically selected based on the parameters provided in the launch file.

The maps created by the SLAM toolbox are saved in the scout_nav2/maps folder. Maps can have different formats:

• pgm and yaml files for 2D maps created using the SLAM toolbox and deserialized using the map_server node, so that the map can be used by the AMCL localization algorithm. The pgm file is a grayscale image of the map, while the yaml file contains the metadata of the map.
• data and posegraph files for 2D maps created using the SLAM toolbox and serialized using the slam_toolbox node, so that the map can be used by the SLAM toolbox algorithm for efficient localization. These files are binary files and can be visualized using the slam_toolbox node.

To launch the simulation of the robot in Gazebo, along with NAV2, run the following command:

$ ros2 launch agilex_scout simulate_control_gazebo.launch.py lidar_type:=<3d|2d> rviz:=<true|false>

$ ros2 launch scout_nav2 nav2.launch.py simulation:=true slam:=<true|false> localization:=<amcl|slam_toolbox>

The first command will launch the simulation of the robot in Gazebo, with the 3D lidar sensor mounted on top, and the RViz GUI (optional). The second command will launch the navigation stack with the specified parameters.

Parameters for simulation launch file:

• lidar_type: 3d for a 3D lidar (pointcloud2), 2d for a 2D lidar (laserscan)
• rviz: true if launching rviz, false if launching only the gazebo simulation

To launch the real scout robot with the lidar sensor and the pointcloud-to-laserscan conversion node, run the following command:

$ ros2 launch agilex_scout scout_robot_lidar.launch.py 

$ ros2 launch scout_nav2 nav2.launch.py simulation:=false slam:=<true|false> localization:=<amcl|slam_toolbox>

Parameters:

• slam: True if you want to use SLAM toolbox for map creation, False if you want to do localization + navigation with the map already created
• simulation: true if running in simulation with gazebo, false if launching the real AgileX Scout robot with real sensors
• localization: choose the localization algorithm, among amcl and slam_toolbox

The repository is organized in different packages:

1. agilex_scout: contains the URDF description of the AgileX Scout robot, the Gazebo simulation configuration description files, and the launch files for the simulation and the real robot control.
2. aws-robomaker-small-warehouse-world: contains the Gazebo world files for the AWS RoboMaker Small Warehouse World.
3. scout_nav2: contains the configuration files and launch files for the navigation of the AgileX Scout robot with Nav2.
   • launch folder: contains the main launch file for the navigation stack.
   • maps folder: contains the maps created by the SLAM toolbox.
   • params folder: contains the YAML configuration files for the navigation stack.
   • rviz folder: contains the RViz configuration files for the visualization of the robot and the navigation stack operation.
4. spatio-temporal-voxel-layer: contains the spatio-temporal voxel layer plugin for the costmap.
5. nav2_bringup_custom: contains the custom version of the nav2_bringup package, which is a fork of the original package. The custom version is needed to allow the use of the nav2_collision_monitor node, which is not available in the original package, and to allow the use of other custom plugins for the navigation stack.