---
title: "πR²：反应式实时流策略——让通用操纵策略在闭环中高频重规划"
source: "arXiv AI/CL/LG（论文）"
source_tier: "t2"
url: "https://arxiv.org/abs/2607.26055v1"
permalink: "https://feed.astralor.com/items/itm_83b67ae1bd309a434b1c0072"
published: "2026-07-28T17:59:31.000Z"
score: 78
channel: "ai"
tags: ["GR00T N1.7","反应式实时控制","扩散强制","机器人操纵","流策略"]
---

# πR²：反应式实时流策略——让通用操纵策略在闭环中高频重规划

## AI 点评

> 我认为这项工作的亮点在于巧妙利用扩散强制的噪声调度实现了条件分离，让大模型在保留表达能力的同时具备实时响应能力。对从事机器人操纵策略、实时控制系统研究的学者或工程师尤其值得一看，它提供了一个即插即用的改进方案，有望推动通用操纵策略在动态环境中的落地。

## 速览

通用操纵策略通常基于大型预训练骨干网络构建动作分块流策略，但该方式开环执行，无法对中途到达的感官输入做出反应，且重规划频率受限于感知-动作管线的延迟。本文提出πR²，利用扩散强制的逐位置噪声调度，将条件分为快通道（本体感觉）和慢通道（视觉语言特征），并设计延迟自适应流调度，使策略在保持大模型和多模态预测能力的同时实现反应式实时控制。在xArm6+XHand平台上，πR²使闭环重规划速度提升约4倍（约25Hz），成功率在仿真和真实环境中分别提高23%和30%。

## 关键点

- 提出将条件解耦为快通道（本体感觉）和慢通道（视觉语言特征），使策略在动作块内对自身状态即时反应，同时容忍视觉特征延迟。
- 设计延迟自适应流调度，将飞行中动作视为修复条件，每次调用仅需一步去噪，单一模型可适应不同硬件延迟。
- 在GR00T-N1.7上微调实现，xArm6+XHand平台闭环重规划频率提升至约25Hz，成功率大幅提升。

## 全文

Generalist manipulation policies increasingly take the form of action-chunking flow policies built on large pretrained backbones. Such chunks run open-loop, so the policy cannot react to sensory input arriving mid-execution, sacrificing \emph{reactivity}. Replanning more often would restore it, but the perception-to-action pipeline (a large backbone plus multiple denoising steps) is too slow: this \emph{latency} forbids frequent replanning and leaves committed actions stale, making such policies ill-suited for dynamic, closed-loop control. We present $π\mathbf{R}^2$, which makes these policies reactive and real-time while retaining large backbones, expressive multi-modal policies, and multi-action prediction. Built on the per-position noise schedule of diffusion forcing, $π\mathbf{R}^2$ contributes two ideas. First, it splits conditioning into a fast channel (proprioception, fresh every tick) and an asynchronously updated slow channel (vision-language features), so the policy reacts to proprioception within a chunk while tolerating stale vision. Second, a latency-adaptive flow schedule treats in-flight actions as inpainting conditioning and emits actions in one denoising step per call, letting one trained model adapt to varying hardware latency. Requiring minimal modification to existing architectures, $π\mathbf{R}^2$ can be finetuned from a pretrained policy: applied to GR00T-N1.7 on a real xArm6+XHand platform, it replans closed-loop roughly $4\times$ faster than the base policy (~$25$Hz on an A5000 GPU), acting on a fresh observation every $40$ms. Across simulation and real-world manipulation tasks, $π\mathbf{R}^2$ improves the success rate by up to $23\%$ in simulation and $30\%$ in the real world over the strongest baseline. Project page: https://pi-r2-flow.github.io/

---

引用前请核对原文;本文由星际流动整理:https://feed.astralor.com/items/itm_83b67ae1bd309a434b1c0072