Partial-Observation Localization State Simulation Dataset for Autonomous Underwater Vehicles
- Records
- 8020
- Size
- 0.3GB
- Format
- JSON
- Provider
MOBIUSI INC- Updated
- v1.02026-11-04
Designed for autonomous underwater vehicles, this dataset simulates observation conditions from inertial navigation, DVL, acoustic positioning, and depth sensors, with structured records of hidden ground-truth states, sensor readings, and their availability. It covers state estimates, missing observations, localization drift, confidence, and state representations to support analysis of state evolution and localization error under partial observability. Corresponding ground-truth and estimated states provide reference data for world model training, sensor fusion research, and simulated localization evaluation.
Overview
Dataset Insights
Preview
Sample Examples
Schema
Technical Specifications
| Field | Type | Description |
|---|---|---|
| file_name | string | File name |
| position_drift | object | Position error and drift-process metrics calculated by comparing the estimated localization state with the hidden ground-truth state. |
| estimated_state | object | An estimate of the robot state inferred from the available partial sensor observations. |
| observation_mask | object | A partial-observation mask generated from the validity flags of the sensor readings. |
| simulation_config | object | Scene, sampling, and coordinate-system settings used to generate the current simulation sample. |
| ground_truth_state | object | The hidden robot state generated by the simulator, used as a reference for localization estimates and error calculations. |
| sensor_observations | object | Raw readings and validity flags from inertial navigation, DVL, acoustic positioning, and depth sensors at the current time step. |
| state_representation | array | A structured state representation vector encoded from hidden-state evolution and partial-observation information for world model learning. |
| localization_confidence | number | A localization confidence score calculated from observation completeness, localization covariance, and sensor quality. |
| observation_completeness | number | The proportion of valid sensor observations at the current time step relative to the total number of sensors. |
Trust
Compliance Statement
Authorization Type
Proprietary - Commercial AI Training License (No Redistribution)
Commercial Use
Requires exclusive subscription or authorization contract (monthly or per-invocation charging)
Privacy and Anonymization
No PII, no real company names, simulated scenarios follow industry standards
Compliance System
Compliant with China's Data Security Law / EU GDPR / supports enterprise data access logs
FAQ
Frequently Asked Questions
- Which sensors and state information are included?
- The dataset simulates inertial navigation, DVL, acoustic positioning, and depth sensor readings. Structured information includes hidden ground-truth states, observation availability masks, estimated states, drift metrics, confidence, and state representations.
- Does the dataset contain real underwater recordings?
- This is a simulation dataset. It provides simulated observations and their corresponding hidden ground-truth states for algorithm development and controlled evaluation; simulated records should not be treated as real sea-trial data.
- Which training and evaluation tasks does it support?
- It supports state representation and estimation under partial observability, localization drift modeling, sensor fusion, and world model training. It can also be used to compare localization performance under different observation availability conditions.
- How can I obtain the data, and what are the licensing terms?
- Check the current product page for available delivery methods, dataset size, format, and licensing terms. Specific access conditions and permitted uses are subject to the information published on the page or the sales agreement.
Can't find the data you need?
Post a request and let data providers reach out to you.
Get this Dataset
High-quality data, ready for your next AI breakthrough.
Cite this Work
@dataset{Mobiusi2026,
title={Partial-Observation Localization State Simulation Dataset for Autonomous Underwater Vehicles},
author={MOBIUSI INC},
year={2026},
url={https://www.mobiusi.com/datasets/0e4625a8db2f5307027cdf78b75f9a8d?dataset_scene_cate_type=8},
urldate={2026-10-28},
keywords={autonomous underwater vehicle dataset,underwater localization simulation,partial observation states,inertial navigation data,DVL observation data,acoustic positioning data,sensor fusion,world model training},
version={1.0}
}Using this in research? Please cite us.