A ROS coordinator and simulation for multi-robot soil properties mapping.
This centralised coordinator directs a team of robots in mapping geospatial data using Ordinary Kriging (OK) interpolation. The kriging variance, a metric of the uncertainty of the interpolation's predictions, is used to decide where robots should sample, and informs allocation of these sampling tasks to the robots.
A simulation for demonstration and testing of this soil mapping multi-robot system, using the move_base_abstract simulator, is included alongside the coordinator in this repository. The simulation re-samples soil compaction data recorded at the University of Lincoln's Riseholme agricultural campus published in the dataset of Fentanes et al.
With your working directory as the src directory of a catkin workspace, clone the repo using:
git clone https://github.com/laurencejbelliott/ROS_multi-robot_soil_mapping_sim.git --recursive
Then navigate to the root directory of the workspace and install the package's dependencies with:
rosdep install --from-paths src --ignore-src -r -y
Finally, build the package with catkin build.
To run an example sampling mission use the command:
rosrun sim_soil_mapping_mrs run_single_trial.py
To replicate the soil compaction mapping trials run for the ICRA '25 submission use the command:
rosrun sim_soil_mapping_mrs run_bid_func_trials.py
Please note this will take a long time to complete, and the data produced from these trials will require a significant amount of free storage space.