Japan Open Driving Dataset Sample: 100+ Hours of Tokyo Driving Data in nuScenes Format

Tokyo, Japan is the geographic scope of this sample subset of a large-scale autonomous driving dataset. The data comprises over 100 hours of driving data, is stored in nuScenes format, and was authored by turing-motors. The dataset page was last updated on 2026-03-27.

Use Cases

• Training autonomous vehicle perception systems based on the included sensor data and 3D annotations.
• Developing Vision-Language-Models (VLMs) based on the virtual captioned data mentioned in the description.
• Researching Vision-Language-Action (VLA) models for driving tasks based on the multimodal data.
• Benchmarking 3D object detection and tracking algorithms using the nuScenes-formatted data.

Strengths

• Contains over 100 hours of driving data, indicating a substantial temporal scope.
• Includes multiple data modalities such as sensor data, 3D annotations, and virtual captioned data.
• Uses the standardized nuScenes format, which likely facilitates compatibility with existing toolkits.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count and file size are unknown, which may limit suitability assessment.

Provenance

Source

turing-motors

Collection Method

Collected from driving in Tokyo, Japan.

Time Range

null

Freshness

Last updated 2026 03 27 03 11 04; freshness should be verified.

Geography

Tokyo, Japan