Home

Awesome

IBISCape: A Simulated Benchmark for multi-modal SLAM Systems Evaluation in Large-scale Dynamic Environments

<p align="center"> <img src="IBISCape_github.gif" alt="Video to Events" width="800"/> </p>

Citation

Are you interested to use IBISCape data acquisition frameworks, or to evaluate your new SLAM system on IBISCape benchmark in an academic work ! Thanks for citing the following paper:

@Article{IBISCape22,
author={Soliman, Abanob and Bonardi, Fabien and Sidib{\'e}, D{\'e}sir{\'e} and Bouchafa, Samia},
title={{IBISCape}: A Simulated Benchmark for multi-modal {SLAM} Systems Evaluation in Large-scale Dynamic Environments},
journal={Journal of Intelligent {\&} Robotic Systems},
year={2022},
month={Oct},
day={19},
volume={106},
number={3},
pages={53},
issn={1573-0409},
doi={10.1007/s10846-022-01753-7},
url={https://doi.org/10.1007/s10846-022-01753-7}
}

To run any of our acquisition frameworks, make sure you have:

IBISCape frameworks are tested on CARLA versions: 0.9.10, 0.9.10.1, and best performance with 0.9.11.

Calibration Acquisition frameworks

$ python3 framework_name.py

The calibration data acquisition is done manually, where the user can collect sequences with all the movements specific to his application.

VI-SLAM Acquisition frameworks

$ python3 framework_name.py

The VIO-SLAM data acquisition is done using CARLA autopilot, following its traffic manager to perfectly simulate the real world.

IBISCape ROS_tools