VARP: A Visual Adaptive Rearrangement Planner for Non-Monotone Object Rearrangement Planning
VARP: A Visual Adaptive Rearrangement Planner for Non-Monotone Object Rearrangement Planning
This research presents VARP, a visual adaptive rearrangement planner that learns non-monotone object rearrangement directly from visual state representations. The system predicts pick-and-place actions using current and goal visual states, incorporating an Action Refinement Strategy (ARS) to enhance feasibility and a Progress Enforcement Strategy (PES) to prevent stagnation. VARP adapts online to disturbances without replanning or combinatorial search. In challenging benchmarks, it achieves a 98.3% success rate with 0.4 seconds planning time per task, maintains 97.5% success under execution-time disturbances, and reaches 94.5% under continuous severe disturbances, while generalizing to unseen object counts.
The paper introduces VARP, a learning-based planner for non-monotone object rearrangement that adapts online to disturbances without combinatorial search, reporting high success rates on benchmarks under various disturbance conditions.
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.
Reported numbers
planning time per task
0.4
View original evidence
rom the latest visual state representations, VARP adapts online to disturbances without replanning from scratch or performing online combinatorial search. In challenging rearrangement benchmarks, VARP achieves a 98.3\% success rate with only 0.4 s planning time per task, 97.5\% success under execution-time disturbances, and 94.5\% under continuous severe disturbances, while generalizing effectively to unseen object c
Open source S4
- paper: not_reported
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
rom the latest visual state representations, VARP adapts online to disturbances without replanning from scratch or performing online combinatorial search. In challenging rearrangement benchmarks, VARP achieves a 98.3\% success rate with only 0.4 s planning time per task, 97.5\% success under execution-time disturbances, and 94.5\% under continuous severe disturbances, while generalizing effectively to unseen object c
Open source S4
Source：arXiv Robotics — research abstracts · arxiv.org