This code was used for the generation of the dataset CARLASENCES, paper

Andreas Kloukiniotis, Andreas Papandreou , Christos Anagnostopoulos, Aris Lalos, Petros Kapsalas , Duongvan Nguyen, Konstantinos Moustakas,”CarlaScenes: A synthetic dataset for odometry in autonomous driving”, CVPR 2022 Workshop on Autonomous Driving

How to access the dataset

The IP of the FTP server to download the dataset is ftp://195.251.58.20:60521

To generate data from carla

  1. Run python multi_data_generator.py

To convert carla data to bag files

  1. Install conda

  2. conda env create --name $conda name$ -f carla_to_bag/env.yml

  3. Replace line 258 from /home/andreas/.local/lib/python2.7/site-packages/pykitti.raw.py:

    • t = dt.datetime.strptime(line[:-4], '%Y-%m-%d %H:%M:%S.%f') -> t = line
  4. Type python carla_to_bag/carla_to_bag.py -h for the documentation

    • Example : python carla_to_bag/carla_to_bag.py -d path_to_ego/../Town03_15_09_2021_14_12_09_to_keep/ego0/ -t 2021_09_15 -r 0003 -f 10
  5. To align poses between vloam and ground truth from carla data type:

    • python carla_to_bag/align_poses.py -h for the documentation
  6. To evaluate results using evo, move LO1.txt, VO1.txt and MO1.txt and aligned_poses from vloam ros results to evo folder and type:

    • evo_traj kitti MO1.txt VO1.txt LO1.txt --ref=aligned_poses.txt -p --plot_mode=xyz -as

ScenarioRunner

To run the files for ScenarioRunner follow [https://carla-scenariorunner.readthedocs.io/en/latest/]

References

  1. Carla Simular [https://github.com/carla-simulator/carla]
  2. Pykitti [https://github.com/utiasSTARS/pykitti]

Related Publications

  1. Countering Adversarial Attacks on Autonomous Vehicles Using Denoising Techniques: A Review
  1. Deep multi-modal data analysis and fusion for robust scene understanding in CAVs