VLA-ULAP: Interleaving Cloud VLA Calls with Ultra-Lightweight Local Action Prediction at the Edge

📄 arXiv: 2609.18663v1 📥 PDF

作者: Deyu Cao, Ryuji Oi, Kosuke Matsushima, Yuxuan Pan, Ziheng Wang, Daichi Fujiki, Atsutake Kosuge

分类: cs.RO, cs.LG

发布日期: 2026-09-16

备注: Preprint

💡 一句话要点

提出VLA-ULAP以解决远程推理延迟问题

🎯 匹配领域: 支柱二：RL算法与架构 (RL & Architecture) 支柱九：具身大模型 (Embodied Foundation Models)

关键词: 视觉-语言-动作 超轻量级预测 实时推理 能耗优化 机器人控制 动态任务

📋 核心要点

  1. 现有的亿级参数VLA策略在设备上消耗大量功耗，且远程推理的延迟影响了实时响应。
  2. 本文提出的VLA-ULAP通过结合远程VLA调用与超轻量级本地动作预测器，显著提高了推理效率。
  3. 实验结果显示，VLA-ULAP在多个基准测试中减少了48.8%至76.7%的VLA调用，同时保持了95.0%至97.5%的成功率。

📝 摘要（中文）

随着亿级参数的视觉-语言-动作（VLA）策略对设备功耗的需求增加，远程推理中的通信延迟影响了响应的及时性。为此，本文提出了VLA-ULAP，它将远程VLA调用与超轻量级本地动作预测器（ULAP）交替进行。ULAP仅包含约740万参数，结合当前视图、运动感知和执行的动作历史进行一次性预测。实验结果表明，ULAP在Jetson Orin Nano上每次推理耗时19.9毫秒，能耗0.183焦耳，相比于RTX A6000上的GR00T，推理时间减少了93.0%，能耗降低了99.6%。在多个基准测试中，VLA-ULAP显著减少了VLA调用，同时保持了高达97.5%的基线成功率。

🔬 方法详解

问题定义：本文旨在解决亿级参数VLA策略在设备上高功耗和远程推理延迟的问题。现有方法在实时应用中面临响应时间长和能耗高的挑战。

核心思路：VLA-ULAP的核心思路是将远程VLA调用与本地的超轻量级动作预测相结合，利用本地信息快速做出决策，从而减少对远程推理的依赖。

技术框架：VLA-ULAP的整体架构包括一个冻结的视觉编码器和一个独立训练的ULAP模块。ULAP通过整合当前视图、运动感知和历史动作数据进行一次性预测，避免了对VLA隐藏状态的依赖。

关键创新：VLA-ULAP的主要创新在于其超轻量级设计，ULAP模块的参数量仅为740万，显著降低了推理时间和能耗，与传统的VLA方法相比，具有本质上的效率提升。

关键设计：ULAP的设计包括对当前视图、运动感知和历史动作的整合，采用了特定的损失函数和网络结构，以优化预测的准确性和速度。

🖼️ 关键图片

fig_0
fig_1

📊 实验亮点

实验结果显示，VLA-ULAP在Jetson Orin Nano上每次推理耗时19.9毫秒，能耗0.183焦耳，相比于RTX A6000上的GR00T，推理时间减少了93.0%，能耗降低了99.6%。在多个基准测试中，VLA-ULAP减少了48.8%至76.7%的VLA调用，同时保持了95.0%至97.5%的成功率。

🎯 应用场景

VLA-ULAP的研究成果在机器人控制、自动驾驶和智能监控等领域具有广泛的应用潜力。通过减少推理延迟和能耗，该方法能够提升实时决策的效率，适用于需要快速响应的动态任务场景。未来，该技术可能推动更多智能设备的普及与应用。

📄 摘要（原文）

Billion-parameter vision--language--action (VLA) policies demand substantial onboard power, while communication delays in remote inference hinder timely responses. We propose VLA-ULAP, which interleaves remote VLA calls with an Ultra-Lightweight Local Action Predictor (ULAP). With approximately 7.4M parameters including the frozen vision encoder, ULAP combines current views, proprioception, and executed action history to predict chunks in one pass. Trained independently, it requires no VLA hidden states, online verification, or server round trips. On Jetson Orin Nano, ULAP takes 19.9 ms and 0.183 J per inference, compared with 284.3 ms and 50.55 J for GR00T on RTX A6000. Across three simulated base-policy/benchmark pairs, selected operating points remove 48.8--76.7\% of VLA calls while retaining 95.0--97.5\% of the baseline success rate. Against local VLA-acceleration alternatives on VLA-JEPA, ULAP uses an estimated 49.2\% less inference time and 51.0\% less GPU energy per successful episode than ACT at comparable success rates, and 77.1\% less time and 79.9\% less energy than SP-VLA at equal success rates. Physical SO-101 experiments retain 95.2--100\% of the baseline success rate across seen and held-out placements while reducing inference time by an estimated 47.9--58.0\% and inference-device energy by 52.1--62.5\%, based on successful-episode call counts and measured device costs. Faster responses also improve dynamic-task success rates: in latency-aware LIBERO-Safety simulation, VLA-ULAP exceeds $π_{0.5}$ by 11.0 and 15.5 percentage points on two tasks while approximately halving VLA calls.