The Intelligent Systems, Automation and Robotics Laboratory

Unreal Dataset #2

We present a simulated dataset collected with Unreal Engine 4 (UE4), a graphics engine with high degree of photorealism, with the MAV dynamics simulated by AirSim. Their combination makes it possible to collect sequences that are very realistic and, therefore, relevant to investigate the performance of VIO/SLAM algorithms.

The simulated drone is equipped with a stereo camera, recording at 20 FPS with 640 × 480 resolution, and an IMU with a 200 Hz sampling rate. We record 24 sequences in six different simulated environments, four sequences each. Three of these environments simulate outdoor scenarios (i.e., Abandoned Factory, City Park, and Urban City), while the remaining ones refer to indoor contexts (i.e., Office Building, Office Space, and Indoor Factory). We provide 4 sequences for each environment.

In the following section, we provide some examples of the indoor and outdoor simulated environments employed for the data collection.

Available data includes:
– Rosbag (IMU, GPS data, images of both the right and left cameras, and cameras information);
– Position ground-truth in TUM format;
– Calibration files.

The dataset is available at the following link:

For a benchmark of VO-VIO approaches on the Unreal Dataset, refer to the following publication:

If you use Unreal Indoor Outdoor Dataset in an academic work, please cite:

@article{legittimo23benchmark, AUTHOR = {Legittimo, Marco and Felicioni, Simone and Bagni, Fabio and Tagliavini, Andrea and Dionigi, Alberto and Gatti, Francesco and Verucchi, Micaela and Costante, Gabriele and Bertogna, Marko}, TITLE = {A benchmark analysis of data‐driven and geometric approaches for robot ego‐motion estimation}, YEAR = {2023}, JOURNAL = {Journal of Field Robotics} }


Indoor Environments

Indoor Factory

Office Building

Office Space

Outdoor Environments

Abandoned Factory

City Park

Urban City