README.md

metadata

license: cc-by-4.0
task_categories:
  - time-series-forecasting
language:
  - en
tags:
  - gnss
  - imu
  - sensor-fusion
  - time-series
pretty_name: 'GINE: GNSS-IMU Next Estimation Dataset'
size_categories:
  - 100K<n<1M

model: https://huggingface.co/ISeeTheFuture/GINE-0.5

train code: https://www.kaggle.com/code/edleedlee/gine-0-5

Dataset Summary

A time-series dataset for predicting the next GNSS position using past GNSS and IMU data. The data was collected from vehicle driving logs and is intended for navigation accuracy improvement and GNSS correction research.

Supported Tasks

• time-series-forecasting (GNSS correction / prediction)

Data Fields

• timestamp (float64) — Time in seconds since the start of recording.

• latitude (float64) — GNSS latitude in decimal degrees.

• longitude (float64) — GNSS longitude in decimal degrees.

• altitude (float64) — GNSS altitude in meters.

• accelerometer_x (float64) — Linear acceleration along the X-axis (m/s²).

• accelerometer_y (float64) — Linear acceleration along the Y-axis (m/s²).

• accelerometer_z (float64) — Linear acceleration along the Z-axis (m/s²).

• gyroscope_x (float64) — Angular velocity around the X-axis (rad/s).

• gyroscope_y (float64) — Angular velocity around the Y-axis (rad/s).

• gyroscope_z (float64) — Angular velocity around the Z-axis (rad/s).

• compass (float64) — Magnetic heading in radians.

• target_latitude (float64) — Ground truth latitude for the next GNSS point.

• target_longitude (float64) — Ground truth longitude for the next GNSS point.

• target_altitude (float64) — Ground truth altitude for the next GNSS point.

Source & Collection

Details about the data collection equipment, driving environments, and sampling frequency should be specified here.

Licensing

This dataset is released under the CC BY 4.0 license. Attribution to the original author ISeeTheFuture is required.