MINOS: Multimodal Indoor Simulator

MINOS is a simulator designed to support the development of multisensory models for goal-directed navigation in complex indoor environments. MINOS leverages large datasets of complex 3D environments and supports flexible configuration of multimodal sensor suites.

 

Demos

MINOS provides both a Python API for training end-to-end navigation agents, and a web client API for crowdsourcing, visualization and debugging.

Try our live web demos:

Use the I/J/K/L keys to step and turn, the arrow keys to strafe and look up or down, the R key to respawn, and the O key to update the agent observations panel on the right.

Download

The MINOS source code is available at https://github.com/minosworld/minos.

For synthetic scenes, we use environments from SUNCG, a dataset with over 45,000 scenes. Please request access to the SUNCG dataset using this form and indicate that you will use the dataset with MINOS.

For reconstructed scenes, we use environments from Matterport3D, a dataset with 90 multi-floor houses containing 2000 room regions. Please sign the terms of use for Matterport3D and send them to matterport3d@googlegroups.com, indicating use with MINOS, to request access to the dataset. Then call the provided download script with the arguments --task_data minos -o .

Paper

If you use MINOS please cite our technical report:

@article{savva2017minos,
  author = {Manolis Savva and Angel X. Chang and Alexey Dosovitskiy and Thomas Funkhouser and Vladlen Koltun},
  title = {{MINOS}: Multimodal Indoor Simulator for Navigation in Complex Environments},
  journal = {arXiv:1712.03931},
  year = {2017}
}