arXiv cs.AI· Kaijun Zhou, Zhiyang Li, Le Chen, Jinyu Gu·· 13 小时前AI 评分32
rMuscle：面向高效 VLA 模型推理的机器人肌肉记忆框架
rMuscle: Robotic Muscle Memory for Efficient Vision-Language-Action Model Inference
AI 导读
rMuscle 是一个受人类肌肉记忆启发的实时 VLA 推理框架，通过双阶段肌肉记忆缓存复用跨执行相似性，在 RTX 4090 和 Jetson Thor 上实现 1.29-1.42 倍加速，并在 LIBERO、RoboTwin 及真实机器人操作任务中保持原有成功率。
正文 · 原文
Abstract:Factory work is a promising early scenario for embodied AI: assigning repetitive manual jobs to robots has clear economic payoff, and a structured station keeps the jobs tractable for current policies. Vision-Language-Action (VLA) models now dominate as the policy paradigm for these robots. The inference latency of VLA models directly affects robot responsiveness and motion smoothness. However, existing VLA inference frameworks do not fully exploit the characteristics of embodied workloads or account for the distinct bottlenecks across different stages of VLA inference.
In this paper, we first characterize embodied workloads and identify substantial task similarity across repeated robot executions. We further find that such similarity extends beyond observations and action trajectories to internal model states. Drawing on these observations, we present rMuscle, a real-time VLA inference framework inspired by human muscle memory. It exploits cross-execution similarity through a dual-phase muscle-memory cache. The Context Cache reuses visual-token outputs to reduce computation, while the Action Cache reuses neuron activation patterns to reduce weight accesses. We keep both the cache memory footprint and access overhead low through online cache recomputation, sliding-window cache retrieval, and mask sharing across consecutive denoising steps. rMuscle achieves 1.29-1.42X speedup on RTX 4090 and Jetson Thor across LIBERO, RoboTwin, and physical manipulation tasks, while maintaining the original success rates on real-world robots.
| Subjects: | Robotics (cs.RO); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2609.19104 [cs.RO] |
| (or arXiv:2609.19104v3 [cs.RO] for this version) | |
| https://doi.org/10.48550/arXiv.2609.19104 arXiv-issued DOI via DataCite |
Submission history
From: Kai Jun Zhou [view email]
[v1]
Wed, 16 Sep 2026 17:34:43 UTC (18,476 KB)
[v2]
Wed, 30 Sep 2026 02:27:49 UTC (18,431 KB)
[v3]
Wed, 21 Oct 2026 07:08:08 UTC (18,431 KB)
来源：arXiv cs.AI · arxiv.org