arXiv Robotics — research abstracts· Manav Vora, Gokul Puthumanaillam, Aditya Penumarti, Jane Shin, Melkior Ornik·· 6 days agoEditorial score65
Akule: Fast and Scalable Multi-Robot Motion Planning via Sparse Interaction Diffusion
Akule: Fast and Scalable Multi-Robot Motion Planning via Sparse Interaction Diffusion
Summary
Akule is a sparse interaction diffusion framework for multi-robot motion planning that combines a frozen single-robot prior with learned modules for pairwise correction, contextual selection, and effect composition. It achieves up to 25× speedups and supports planning for up to 256 robots across diverse benchmarks.
Source: arXiv Robotics — research abstracts · Read original article ↗
Loading article text…
Source：arXiv Robotics — research abstracts · arxiv.org