This page contains the accompanying data for the CVAE Exploration Planning project. The code to train our models and run a planner is available at https://github.com/ethz-asl/cvae_exploration_planning. The paper preprint is available on ArXiv:

Schmid, Lukas, and Ni, Chao, et al. “Fast and Compute-efficient Sampling-based Local Exploration Planning via Distribution Learning”, in IEEE Robotics and Automation Letters (RA-L), 2022.

The following files are provided (# short description). Additional processing and use tools are available on github.

Data:

• CVAE_dataset.npy # Numpy data containing all our local maps with the best actions identified by the teacher planner as supervision signal. Used to train or CVAE models.
• CNN_dataset.npy # Numpy data containing local maps with random actions used to train the CNN models.
• test_worlds.zip # Directory containing all separate test worlds used in our experiments.
  • <environment>_<no>.p # Pickle file containing the test world.
  • <environment>_<no>.txt #Text file containing the initial position as [x, y, yaw] used in our experiments.

Data Repository

CVAE_dataset.npy (326.4 MB)

CNN_dataset.npy (1.0 GB)

test_worlds.zip (56.5 KB)