Reproduce “Narrow Passage Path Planning using Collision Constraint Interpolation” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Minji Lee, Jeongmin Lee, Dongjun Lee
A framework for robot path planning in narrow passages introduces collision constraint interpolation, systematically relaxing and tightening constraints to avoid deep penetration and local minima. It achieves significantly higher success rates and often faster planning times compared to baseline methods in complex manipulation and nonholonomic navigation tasks.
Key insights
- Gradually transitioning collision constraints from a relaxed 'expanded' free space to the full environment guides optimization to avoid local minima and deep penetration into obstacles.
- Smoothing the surfaces of interpolated objects using a shaping function on their SDFs maintains consistent contact normals, facilitating stable gradient-based optimization.
- Introducing new environmental objects in 'leaf sets' preserves the homotopy of the free space, ensuring that the path can be continuously deformed to remain feasible as constraints tighten.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.