MSaalaamaa/marine_vla_dataset_v1.1

Marine VLA Dataset — v1.1 (Re-collected)

This dataset contains 16,748 frames (12,070 original + 4,678 supplementary) re-collected in the Mangalia harbor Gazebo simulation using the fixed expert controller. It replaces the original v1 dataset which was corrupted by a yaw-spin bug.

Built with opencode + DeepSeek V4 Flash — every component of this project was developed using opencode CLI with DeepSeek V4 Flash as the underlying model.

Why v1.1?

The v1 dataset was collected with zero hydrodynamic damping in the vessel SDF, causing uncontrolled yaw rotation. Labels oscillated between PORT_15 / STARBOARD_15 / PROCEED as the vessel wobbled. A SmolVLA 450M classifier trained on v1 achieved only 25.7% accuracy (see training notebook, Cell 9).

v1.1 fixes:

• Added angular velocity decay (<velocity_decay angular=0.01>) to the vessel SDF
• Added proportional heading-hold during PROCEED to prevent oscillation
• Removed lateral force (Fy) from turns — uses only differential torque
• Corrected port/starboard sign convention

The benchmark achieves 7/7 phases PASS: 5 m/s forward speed, ~30° turns, yaw damping < 2°/s.

Label Vocabulary

ID	Label	Description
0	PROCEED	Continue on heading at cruise speed
1	PORT_15	Turn 15° to port (left)
2	STARBOARD_15	Turn 15° to starboard (right)
3	REDUCE_SPEED	Slow to minimum steerage
4	STOP	Stop engines
5	ASTERN	Reverse thrust
6	AVOID_OBSTACLE	Emergency evasive maneuver
7	DOCK_APPROACH	Orient and approach dock

Data Format

LeRobot-compatible format with image + label pairs:

• observation.images.camera_0: 640×480 RGB, ~50KB JPG
• action.goal_label: integer label ID 0-7

Collection Pipeline

1. ROS Noetic + Gazebo 11 with Mangalia harbor world
2. Custom vessel SDF with camera + force plugin
3. Scenario loader randomizes start pose, waypoints, obstacles
4. Fixed expert controller drives vessel + publishes labels
5. Dataset recorder saves synchronized image + label pairs
6. Converter produces LeRobot-format JSON

Label Distribution (v1.1 + supplement)

Label	Original	After Supplement	Improvement
PROCEED	4,449	6,590	+48%
PORT_15	3,015	3,413	+13%
STARBOARD_15	2,694	3,040	+13%
REDUCE_SPEED	587	1,001	+71%
STOP	167	298	+78%
ASTERN	757	1,364	+80%
AVOID_OBSTACLE	117	406	+247%
DOCK_APPROACH	284	636	+124%

Max/min ratio improved from 38x → 22x.

Repository Structure

• marine_vla_dataset_v1.1.tar.gz — original dataset archive (~95MB, 43 episodes, 12,070 frames)
• marine_vla_dataset_supplement_v1.1.tar.gz — supplementary episodes targeting underrepresented labels (~33MB, 15 episodes, 4,678 frames)
• 58 episodes total, 30s duration each
• 8 scenarios: open transit, port/starboard turns, obstacle avoidance, speed reduction, emergency stop, docking, mixed