Argoverse: Self-Driving Vehicle Datasets with Detailed Maps

Argoverse provides public datasets for testing and teaching self-driving vehicles how to understand the world. The collection includes motion forecasting, tracking, and lidar sensor data from Argoverse 1 and 2, supported by detailed maps. It is hosted on AWS Open Data and maintained by the Argoverse organization.

Use Cases

• Train motion forecasting models based on vehicle trajectory data.
• Develop and test 3D object tracking algorithms based on sensor data.
• Validate perception systems using synchronized lidar and camera data.
• Research map change detection for autonomous vehicle localization.

Strengths

• Includes multiple specialized datasets (Argoverse 1, Argoverse 2, TbV) for different autonomous vehicle tasks.
• Data is supported by detailed maps, which is a key feature for realistic simulation and testing.
• Hosted on a major public cloud platform (AWS), suggesting reliable access.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count and dataset size are unknown, which may limit suitability assessment.
• Last update date is unknown; freshness unverified.

Provenance

Source

Argoverse organization

Collection Method

Likely collected from sensors on autonomous vehicle platforms.

Time Range

null

Freshness

Last updated: unknown

Geography

null