Visualized in Foxglove: Offroad AV Tartan Drive 2.0 by Airlab Dataset
The TartanDrive 2.0 dataset, visualized using Foxglove, was designed to support advanced research in off-road autonomous driving. This dataset provides extensive multimodal data collected through diverse and challenging terrains in western Pennsylvania, USA. It facilitates a range of tasks in perception, planning, and control for autonomous vehicles.
The dataset was captured using a Yamaha Viking All-Terrain Vehicle (ATV), equipped with two Velodyne VLP-32 lidar sensors, one Livox Mid-70 lidar sensor (mounted under the MultiSense camera), a Carnegie Robotics MultiSense S21 providing stereo images and IMU, a NovAtel PROPAK-V3-RT2i GNSS delivering IMU data and fused GPS data that’s providing pose estimates –all streaming at rates from 10Hz up to 400Hz as the ATV itself drives around at incredibly quick speeds.
The TartanDrive 2.0 dataset was created by the AirLab at Carnegie Mellon University by Matthew Sivaprakasam, Parv Maheshwari, Mateo Guaman Castro, Samuel Triest, Micah Nye, Steve Willits, Andrew Saba, Wenshan Wang, and Sebastian Scherer.
Link to the dataset and more about the project in the comments.
https://theairlab.org/TartanDrive2/