benchmark_lunar

Benchmark framework for lunar navigation

This is a [ROS2] humble package developed for benchmark framework for lunar navigation. The software is designed for (local) navigation tasks with robots on lunar environment, which are equipped with a pose estimation and distance sensor. The goal of this software is to provide benchmark framework for lunar navigation.

Building

In order to install the benchmark_lunar, clone the latest version from this repository into your workspace and compile the package using ROS2.

cd your_workspace/src
git clone https://github.com/er1c-cartman/benchmark_lunar.git
cd ../
colcon build

Lunar simulator

First, set your model path to

export GAZEBO_MODEL_PATH=$GAZEBO_MODEL_PATH:$HOME/your_workspace/src/scout_v2/scout_gazebo/models

And run the simulation with

 ros2 launch scout_gazebo moon.launch.py

Elevation mapping

Elevation mapping code is originated from ETH Zurich

• Elevation mapping

And run the elevation mapping and traversability estimation with

 ros2 launch elevation_mapping elevation_mapping.launch.py 

To edit parameters of traversability estimation, you can change values in following file

$HOME/your_workspace/src/elevation_mapping/elevation_mapping/config/grid_map_filters.yaml

Navigation

Navigation code is originated from NAV2

• Navigation 2

And run the navigation with

 ros2 launch scout_navigation mppi.launch.py 
 (dwa, teb, purepursuit. You can edit the config file and launch file to run what you want)

Benchmark

Benchmark code is located at

$HOME/your_workspace/src/scout_v2/scout_navigation/result/benchmark.py

The result is saved at separate folder in same location. You can use python or matlab to benchmark the results