Support-Safe Variational Hybrid Filtering for Contact-Mode and Sparse-Law Recovery

2026-05-12

AI Summary

This work addresses the challenge of filter degeneracy caused by hybrid dynamics in contact-rich robotic systems. The authors propose VHYDRO, a method that jointly infers continuous latent states and discrete contact modes by mixing learned proposal distributions with feasible transition laws at each time step, followed by importance weighting. To ensure filtering stability, VHYDRO incorporates a support-safety mechanism that enforces temporal consistency in contact mode segmentation. Notably, it is the first approach capable of successfully identifying sparse port-Hamiltonian models over pure-mode segments. The method maintains filter robustness under severe occlusion and outperforms baseline approaches in contact segmentation on ManiSkill and Sawyer/BridgeData tasks, while accurately recovering active physical terms in known hybrid systems.

Abstract

Contact-rich robot dynamics are hybrid: a single observation can match several latent states and contact regimes (free, impact, stick--slip). A standard amortized filter that places no probability on a feasible contact transition will permanently lose the branch the robot actually follows. We introduce VHYDRO, a variational hybrid dynamics learner that prevents this branch loss. At each step, VHYDRO mixes the learned proposal with a feasible transition law before sampling and importance weighting, ensuring that every transition retained by the model-feasible carrier remains covered. VHYDRO jointly infers a continuous latent state and a discrete contact mode, and fits a sparse port-Hamiltonian law to each recovered regime. On top of this, three guarantees connect: support coverage stabilizes filtering, the stabilized filter concentrates the discrete contact posterior on coherent regimes, and mode-pure segments admit sparse port-Hamiltonian recovery. The recovery error separates cleanly into filtering, derivative, mode-impurity, and physics-residual parts. Three empirical findings track the same mechanism. Under heavy occlusion the support-safe filter stays usable while a non-defensive proposal collapses. On ManiSkill demonstrations and on four Sawyer/BridgeData task families the discrete state forms temporally coherent contact-regime segments that the discrete state yields a stronger joint profile across ARI, change-point F1, and segment purity than post-hoc and mode-free baselines. On hybrid systems with known equations the mode-conditioned sparse fit recovers the active physical terms; purely predictive baselines do not.

Problem

• hybrid dynamics
• contact-rich robotics
• state filtering
• mode recovery
• sparse physical laws

Innovation

• support-safe filtering
• hybrid dynamics
• variational inference
• sparse port-Hamiltonian
• contact mode recovery