BASEPROD Scripts

These scripts were used to prepare BASEPROD: The Bardenas Semi-Desert Planetary Rover Dataset. The code is released under the MIT License.

Authors:

• Levin Gerdes
• Hugo Leblond
• Rául Castilla Arquillo
• Joaquín Ortega Cortés
• Felix Wilting

Supervisor: Carlos J. Pérez del Pulgar

Links

• Paper: https://doi.org/10.1038/s41597-024-03881-1
• Data: https://doi.org/10.57780/esa-xxd1ysw

Requirements

The code was tested with Python 3.10.12. Additional Python requirements are listed in requirements.txt and can be installed in a virtual environment as follows:

python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

Activate the environment with source .venv/bin/activate or deactivate it by invoking deactivate.

Some scripts use additional packages for plotting:

sudo apt install \
texlive \
texlive-latex-extra \
cm-super \
dvipng

The scripts are formatted with black (exact version).

Usage

Download and unzip the (relevant parts of the) dataset to your local drive. For convenience, we recommend creating a shorthand via export BASEPROD=/path/to/baseprod. This step is assumed for the following example commands.

Note

View detailed script usage information via python <script> --help.

Exporting mcap recordings

You can export all data recordings via

python export_logs.py -i ${BASEPROD}/rosbags/ -o ${BASEPROD}/rover_sensors/

This will iterate over all roslogs in that directory and take ages, because we write to disk quite often to avoid running out of memory.

Plotting weather station

To plot weather station data, provide the CSV of the day and the data you are interested in. E.g., to plot all data from the 23rd of July:

python plot_weather.py --file ${BASEPROD}/weather_station/2023-07-23_09-49-44.csv --all

Plotting LIBS spectra

Spectra can be plotted by providing a spectra.csv file and indicating the spectrum of interest. E.g., to plot spectrum 3 of the 2nd measurement at location 1:

python plot_spectrum.py --file ${BASEPROD}/libs_measurements/01/02/spectra.csv --spectrum 3

Data corrections

The following commands compute and apply corrections to different measurements. All input data will be preserved; corrected data can be found in files that contain "CORRECTED" in their filenames.

The Force/Torque corrections need an exported calibration log (passed via -c flag) and an exported input log to correct (-i).

python ft_correction.py -c ${BASEPROD}/rover_sensors/calibration/ft_neutral -i ${BASEPROD}/rover_sensors/2023-07-23_12-52-39

The bogie offset script only needs an exported input log (-i).

python bogie_offset.py -i ${BASEPROD}/rover_sensors/2023-07-23_12-52-39

The gyro offset script needs an exported input log (-i) and can be given a manual offset correction (-c) for the heading in radians.

python gyro_offset.py -i ${BASEPROD}/rover_sensors/2023-07-23_12-52-39 -c -1.25

Tip

We processed most data in batch via the following commands.

export TPATH=/path/to/baseprod/rover_sensors
ls ${TPATH} | grep 2023 | xargs -I {} python ft_correction.py -c ${TPATH}/calibration/ft_neutral -q -i ${TPATH}/{}
ls ${TPATH} | grep 2023 | xargs -I {} python bogie_offset.py -q -i ${TPATH}/{}
ls ${TPATH} | grep 2023 | xargs -I {} python gyro_offset.py -q -i ${TPATH}/{}

Citation

If you use this work, please cite:

Gerdes, L., Wiese, T., Castilla Arquillo, R. et al. BASEPROD: The Bardenas Semi-Desert Planetary Rover Dataset. Sci Data 11, 1054 (2024). https://doi.org/10.1038/s41597-024-03881-1

@article{Baseprod,
        author = {Levin Gerdes and Tim Wiese and Raúl Castilla Arquillo and Laura Bielenberg and Martin Azkarate and Hugo Leblond and Felix Wilting and Joaquín Ortega Cortés and Alberto Bernal and Santiago Palanco and Carlos Pérez del Pulgar},
        doi = {10.1038/s41597-024-03881-1},
        issn = {2052-4463},
        issue = {1},
        journal = {Scientific Data},
        month = {9},
        pages = {1054},
        title = {BASEPROD: The Bardenas Semi-Desert Planetary Rover Dataset},
        volume = {11},
        url = {https://www.nature.com/articles/s41597-024-03881-1},
        year = {2024},
}