WasserMan

Benchmark for Underwater Manipulation Policy Learning.

WasserMan is a simulation benchmark for underwater vehicle–arm manipulation, with contact tasks, water forces and visual policy learning.

Key Features

• 10 tasks — contact, recovery, delivery and underwater intervention; nine have learned-policy evaluations.
• Two robot platforms — BlueROV2 Heavy + Alpha 5 and RexROV2 + Oberon7, plus bimanual valve manipulation.
• Visual baselines — ACT, Diffusion Policy and chunked BC on six tasks; SmolVLA on three tasks, each with three training seeds.
• Underwater physics and control — currents, buoyancy, thruster response, PID/PD comparisons, optional wall/seabed effects and camera imaging.
• Reproducible evaluations — task contracts, splits, checkpoints and recorded outcomes: 3,780 learning/intervention episodes and 270 SmolVLA test episodes.

Installation

Requires native Linux, Python 3.12 and a compatible NVIDIA GPU/driver. The open core uses Isaac Sim 6.1, Isaac Lab and procedural arm geometry.

git clone --branch main https://github.com/dancher00/wasserman.git
cd wasserman
./scripts/bootstrap.sh --asset-profile open-procedural-v1
.venv/bin/python scripts/setup_policy_dependencies.py
.venv/bin/python scripts/setup_training_backbones.py
uv pip install --python .venv/bin/python --no-deps -r research/media-requirements.txt
export WASMAN_ASSET_PROFILE=open-procedural-v1

The installer verifies the scientific runtime files bundled with this checkout. See Installation and Verify installation.

Quick Start

Step an environment with zero actions:

OMP_NUM_THREADS=4 .venv/bin/python scripts/rollout_button_visual.py \
  --mode zero --purpose development --seeds 31000 --steps 16 \
  --output-dir artifacts/installation-zero

Record a development demonstration:

OMP_NUM_THREADS=4 .venv/bin/python scripts/benchmark.py collect \
  --task PressButton --purpose development --seeds 31000 \
  --output-dir artifacts/button-demo

Use a fresh output directory for each run. The 16-step check verifies loading and stepping; it does not measure success. Follow the training guide and frozen protocol for full experiments. No pretrained visual-policy download is needed to start.

Documentation

Guides cover collection, training, evaluation and extending the benchmark. See models and evidence, policy results, including SmolVLA, bimanual studies and Contributing. The website source lives in a separate repository.

Citation

Please cite the WasserMan paper on arXiv:

@misc{belov2026wasserman,
  title = {{WasserMan}: Benchmark for Underwater Manipulation Policy Learning},
  author = {Belov, Danil and Erkhov, Artem and Parsegov, Sergei and Osinenko, Pavel},
  year = {2026},
  eprint = {2610.04536},
  archivePrefix = {arXiv},
  primaryClass = {cs.RO},
  doi = {10.48550/arXiv.2610.04536},
  url = {https://arxiv.org/abs/2610.04536}
}

Machine-readable citation metadata: CITATION.cff.

License

Apache License 2.0. Third-party components retain their own terms; restricted robot CAD and Isaac Sim binaries are not redistributed. See Third-party notices.