Odom Dataset: 17 ROSBag Sessions for Odometry, 34 Minutes Total

A first version ROSBag odometry dataset containing 17 sessions totaling approximately 2077.404 seconds (34 minutes and 37.4 seconds) of data. The repository, created by ly041021, stores raw ROS1 bag files, maintains basic metadata, and provides Parquet index files for Hugging Face compatibility. It was last updated on 2026-05-18 and does not include pre-split training, validation, or test sets, nor exported RGB, depth, infrared, or LiDAR point cloud files.

Use Cases

• Developing odometry algorithms based on raw ROS1 bag files.
• Benchmarking sensor fusion methods using the indexed topic metadata.
• Creating custom data processing pipelines for ROSBag data based on the provided session structure.

Strengths

• Contains 17 distinct data collection sessions.
• Provides a total duration of approximately 2077.404 seconds of sensor data.
• Includes a metadata file with 262 topic statistics records for indexing.

Limitations

• No column-level documentation is provided; field semantics must be inferred after download.
• The dataset is not split into training, validation, or test sets.
• Row count and specific file formats for the core sensor data are unknown.

Provenance

Source

huggingface

Collection Method

Likely collected from robotic platforms using the ROS (Robot Operating System) framework.

Freshness

Last updated 2026-05-18 07:24:50; freshness should be verified.