Synthetic 6-DOF USV Motion Dataset

This repository contains the synthetic 6-degree-of-freedom (6-DOF) vessel motion dataset used in the study:

Epistemic Uncertainty in Short-Horizon Ship Motion Prediction for USVs

This dataset contains the physical specifications, hydrodynamic modeling parameters, and experimental configuration for a small-scale displacement-type Unmanned Surface Vessel (USV). The platform is specifically engineered for high-fidelity maritime research, focusing on active stabilization and predictive digital twin methodologies.

Vessel Specifications

The USV's mass properties and dimensions are defined within a specific mesh reference frame to support numerical replication.

Principal Dimensions

Length Overall (LOA): 1.650 m

Length Between Perpendiculars (LPP/LWL): 1.439 m

Beam (BOA/BWL): 0.800 m

Depth: 0.400 m

Draft: 0.153 m

Mass & Inertial Properties

Displacement: 127.9 kg (Freshwater, $\rho$ = 999 kg/m³)

Center of Gravity (LCG, TCG, VCG): [0.690 m, 0.000 m, 0.047 m]

Radii of Gyration (kxx, kyy, kzz): [0.296 m, 0.389 m, 0.389 m]

Dataset Structure

The dataset consists of frequency-domain hydrodynamic simulations generated using HydroSTAR and post-processed into time-domain motion sequences for machine learning experiments. The full dataset comprises 6,552 simulation cases spanning a structured grid of environmental and operational parameters:

Significant wave height (Hs): 0.25–2.0 m

Peak wave period (Tp): 1.5–6.0 s

Wave encounter direction (β): 0°–180°

Vessel forward speed (U): 0–0.6 m/s

Each simulation has a duration of 10 minutes with a time step of 0.05 s and contains 6-DOF vessel motions.

Files are organized into folders according to vessel forward speed. Each filename encodes the corresponding sea-state parameters. For example:

SS845_Hs2p00_Tp2p00_Beta180_U0p6mps_series.csv

where:

Hs2p00 → significant wave height = 2.00 m

Tp2p00 → peak period = 2.00 s

Beta180 → wave encounter direction = 180°

U0p6mps → forward speed = 0.6 m/s

Each CSV file contains time-series data for 6-DOF vessel motions processed for supervised learning.