HDFlow: Hierarchical Diffusion-Flow Planning for Long-horizon Robotic Assembly
Lists
Only you see this
Abstract
Long-horizon manipulation tasks represent a significant challenge in robotics, demanding both strategic, high-level reasoning and fast, precise, low-level control. While recent advances in generative models have shown promise in generating behavior plans for long-horizon tasks, they often lack a principled framework for hierarchical decomposition and struggle with the computational demands of real-time execution, due to their iterative denoising process. In this work, we introduce $\textbf{Hierarchical Diffusion-Flow}$ ($\texttt{\textbf{HDFlow}}$), a novel hierarchical planning framework that optimally leverages the strengths of $\textit{diffusion}$ and $\textit{rectified flow}$ models. $\texttt{\textbf{HDFlow}}$ employs a high-level diffusion planner to generate sequences of strategic subgoals in a learned latent space, capitalizing on diffusion's powerful exploratory capabilities. These subgoals then guide a low-level rectified flow planner that generates smooth and dense trajectories, exploiting the speed and efficiency of ordinary differential equation (ODE)-based trajectory generation. This hybrid approach synergistically combines the strengths of both models to overcome the limitations of single-paradigm generative planners, enabling robust and efficient long-horizon planning. We evaluate $\texttt{\textbf{HDFlow}}$ on four challenging furniture assembly tasks, where it significantly outperforms state-of-the-art methods. Project website: https://hdflow-page.github.io/
BibTeX
@inproceedings{nandiraju2025hdflow,
author = {Gireesh Nandiraju and Yuanliang Ju and Chaoyi Xu and Weiheng Liu and Yuxuan Wan and He Wang},
title = {{HDFlow: Hierarchical Diffusion-Flow Planning for Long-horizon Robotic Assembly}},
booktitle = {EWM @ NeurIPS},
year = {2025},
url = {https://openreview.net/forum?id=NDGP58PuQX},
}Loading related content...