mlsanityData to 14 Oct 2026

Formal Safety Verification and Refinement for Generative Motion Planners via Certified Local Stabilization

arXiv 2509.19688 ↗First posted 24 Sep 2025cs.RO cs.LG cs.SYcited by 2

Not on the map. Why?

Abstract (arXiv)

We present a method for formal safety verification of learning-based generative motion planners. Generative motion planners (GMPs) offer advantages over traditional planners, but verifying the safety and dynamic feasibility of their outputs is difficult since neural network verification (NNV) tools scale only to a few hundred neurons, while GMPs often contain millions. To preserve GMP expressiveness while enabling verification, our key insight is to imitate the GMP by stabilizing references sampled from the GMP with a small neural tracking controller and then applying NNV to the closed-loop dynamics. This yields reachable sets that rigorously certify closed-loop safety, while the controller enforces dynamic feasibility. Building on this, we construct a library of verified GMP references and deploy them online in a way that imitates the original GMP distribution whenever it is safe to do so, improving safety without retraining. We evaluate across diverse planners, including diffusion, flow matching, and vision-language models, improving safety in simulation (on ground robots and quadcopters) and on hardware (differential-drive robot).

Key facts. Cited by 2 papers in the map (data through 14 Oct 2026). Builds on 27 of its 40 references that are in the map. First posted on arXiv 24 Sep 2025.

Builds on

27 of 40 references.

References · 27

  1. 2020Denoising Diffusion Probabilistic Modelscited by 21,076
  2. 2022Flow Matching for Generative Modelingcited by 5,702
  3. 2018Neural Ordinary Differential Equationscited by 3,779
  4. 2022Is Conditional Generative Modeling all you need for Decision-Making?cited by 516
  5. 2018Efficient Neural Network Robustness Certification with General Activation Functionscited by 445
  6. 2023Motion Planning Diffusion: Learning and Planning of Robot Motions with Diffusion Modelscited by 161
  7. 2021Safe Nonlinear Control Using Robust Neural Lyapunov-Barrier Functionscited by 94
  8. 2023SafeDiffuser: Safe Planning with Diffusion Probabilistic Modelscited by 75
  9. 2024Learning Robotic Manipulation Policies from Point Clouds with Conditional Flow Matchingcited by 65
  10. 2022Adaptive Conformal Prediction for Motion Planning among Dynamic Agentscited by 53
  11. 2021Reachability Analysis of Neural Feedback Loopscited by 41
  12. 2024CoBL-Diffusion: Diffusion-Based Conditional Robot Planning in Dynamic Environments Using Control Barrier and Lyapunov Functionscited by 33
15 more
  1. 2016Funnel Libraries for Real-Time Robust Feedback Motion Planningcited by 29
  2. 2020Planning with Learned Dynamics: Probabilistic Guarantees on Safety and Reachability via Lipschitz Constantscited by 25
  3. 2023Multi-Agent Reachability Calibration with Conformal Predictioncited by 25
  4. 2024DiffuSolve: Diffusion-based Solver for Non-convex Trajectory Optimizationcited by 20
  5. 2022Safe Output Feedback Motion Planning from Images via Learned Perception Modules and Contraction Theorycited by 19
  6. 2022Convex synthesis and verification of control-Lyapunov and barrier functions with input constraintscited by 14
  7. 2025DDAT: Diffusion Policies Enforcing Dynamically Admissible Robot Trajectoriescited by 14
  8. 2021A Simple and Efficient Sampling-based Algorithm for General Reachability Analysiscited by 12
  9. 2022Statistical Safety and Robustness Guarantees for Feedback Motion Planning of Unknown Underactuated Stochastic Systemscited by 11
  10. 2022One-Shot Reachability Analysis of Neural Network Dynamical Systemscited by 10
  11. 2023Efficient Interaction-Aware Interval Analysis of Neural Network Feedback Loopscited by 9
  12. 2023Learning Complex Motion Plans using Neural ODEs with Safety and Stability Guaranteescited by 8
  13. 2016Using Neural Networks to Compute Approximate and Guaranteed Feasible Hamilton-Jacobi-Bellman PDE Solutionscited by 5
  14. 2024Constraint-Aware Refinement for Safety Verification of Neural Feedback Loopscited by 4
  15. 2025Dynamics-aware Diffusion Models for Planning and Controlcited by 4

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 2.

2 papers.

  1. 2026Safe Large-Scale Robust Nonlinear MPC in Milliseconds via Reachability-Constrained System Level Synthesis on the GPUcited by 7
  2. 2026Scalable Data-Driven Reachability Analysis and Control via Koopman Operators with Conformal Coverage Guaranteescited by 1

Similar, not cited

Close in wording, no citation either way

  1. 2026Learning from Demonstrations meets Symbolic Control - A Supervisory Control Architecture for Formal Certification of Learned Motion
  2. 2024Verified Safe Reinforcement Learning for Neural Network Dynamic Models
  3. 2026Let the Dynamics Flow: Stable Flow Matching Dynamical Systems
  4. 2025SafeFlow: Safe Robot Motion Planning with Flow Matching via Control Barrier Functions
  5. 2020GLAS: Global-to-Local Safe Autonomy Synthesis for Multi-Robot Motion Planning with End-to-End Learning
  6. 2026Neuro-Symbolic Safety Guidance for Vision-Language-Action Models via Constrained Flow Matching
  7. 2026Safe Vision Language Action Models via Barrier Enhanced Flow Matching
  8. 2025Generalizing Safety Beyond Collision-Avoidance via Latent-Space Reachability Analysis