Learn-To-Propagate: Distributionally Robust Motion Planning for Nonlinear Uncertain Systems
Learn-To-Propagate: Distributionally Robust Motion Planning for Nonlinear Uncertain Systems
This research presents Learn-To-Propagate Distributionally-Robust Forest (L2P-DR-Forest), a novel sampling-based planning algorithm designed for nonlinear stochastic systems with unknown initial states and process noise distributions. The approach utilizes optimal transport to learn ambiguity sets containing true system distributions with high confidence, propagates these sets through dynamics, and verifies safety. It integrates these techniques into a forest of search trees in the state space while mitigating ambiguity-set growth. The authors prove the algorithm's correctness and probabilistic completeness under certain assumptions. Evaluations on robotic benchmarks demonstrate that L2P-DR-Forest significantly outperforms existing methods while providing formal guarantees with high confidence.
The paper introduces L2P-DR-Forest, a distributionally robust sampling-based planning algorithm for nonlinear stochastic systems with unknown initial states and noise distributions, leveraging optimal transport to learn ambiguity sets and providing formal correctness guarantees; evaluations on robotic benchmarks claim significant performance improvements over existing methods.
Source: arXiv Robotics — research abstracts · Read original article ↗
Loading article text…
What the source reports
Publisher-reported claims, with original evidence. These results have not been independently verified by RoboSignal.
- paper: not_reported
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
arXiv:2611.13863v1 Announce Type: new Abstract: Motion planning for stochastic systems aims to find a sequence of control inputs that drives a system to a goal set while avoiding unsafe states with high probability. Existing approaches, however, are either restricted to linear dynamics with perfect knowledge of the underlying distributions or lack formal correctness guarantees. To address these limitations, we propos
Open source S2
Source：arXiv Robotics — research abstracts · arxiv.org