Metraipolis › archive › dc_e4afcff66cf1

Safe and Robust Neural Policy Learning with Statistical Verification for Sim-to-Real Deployment in Robotics — arXiv:2608.06481

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.06481
License (as archived)
CC0-1.0
Archived
2026-10-21T08:59:28.187Z
SHA-256 of the archived text
937f4a9f22876e00180dea5dad0747fe77dba537f398c575944ab8426072c1d4
Stable URL of this page
https://metraipolis.com/documento/dc_e4afcff66cf1

How to cite

Safe and Robust Neural Policy Learning with Statistical Verification for Sim-to-Real Deployment in Robotics — arXiv:2608.06481 — arXiv (en), https://arxiv.org/abs/2608.06481 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_e4afcff66cf1 (Metraipolis archive dc_e4afcff66cf1, sha256 937f4a9f22876e00180dea5dad0747fe77dba537f398c575944ab8426072c1d4)
BibTeX
@misc{dce4afcff66cf1,
  title = {Safe and Robust Neural Policy Learning with Statistical Verification for Sim-to-Real Deployment in Robotics — arXiv:2608.06481},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.06481},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_e4afcff66cf1; sha256 937f4a9f22876e00180dea5dad0747fe77dba537f398c575944ab8426072c1d4},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.06481. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Synthesizing safe and robust neural controllers in simulation for reliable sim-to-real deployment remains a critical challenge in robotics. Existing learning-based methods typically lack safety and performance guarantees over an explicitly defined operating region, while post-training verification techniques provide no mechanism to refine controllers when safety violations are detected. To bridge this gap, we propose a curriculum-driven framework that tightly integrates scenario-based Evolution Strategy with Statistical Model Checking-based verification in a closed-loop procedure. Starting from a candidate region, our approach co-optimizes policy performance while progressively enlarging its safe operating boundaries. Upon termination, it yields a neural controller together with a region over which safety and performance are statistically verified. Extensive evaluations on Cartpole and 3D Quadrotor benchmarks, showing 6.14x and 224.04x expansions, respectively, of the safe operating region over mathematically certified ones, together with physical experiments under both nominal conditions and severe dynamic perturbations, demonstrate that our learned controllers consistently outperform established control-theoretic and learning-based baselines. Furthermore, we show that the size of the verified region serves as a quantitative indicator of policy quality before deployment. These results establish our framework as an automated pipeline for learning, assessing and deploying safe and robust neural controllers from simulation to reality.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-10 in cs.NE).
Categorie arXiv: Robotics (cs.RO) ; Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Neural and Evolutionary Computing (cs.NE).
Pagina del lavoro: https://arxiv.org/abs/2608.06481 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-21T08:59:27.898Z dalla pagina https://arxiv.org/catchup/cs.NE/2026-08-10?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).