X-Mobility is a dataset for robot mobility training released by NVIDIA in 2024, designed to train world models and action networks using simulated data.

The dataset provides training, validation, and test data for the X-Mobility project, including policy datasets generated using the Isaac Sim simulator. It comprises teacher policy datasets and random action datasets, supporting the pre-training of world models and joint training of world models and action networks, thereby providing high-quality simulation data support for robot mobility research.

Dataset Composition

The dataset primarily consists of the following two core components:

• x_mobility_isaac_sim_nav2_100k.zip: Teacher policy dataset, containing 100k samples, used for joint training of world models and action networks.
• x_mobility_isaac_sim_random_160k.zip: Random action dataset, containing 160k samples, used for pre-training world models without action networks.