arXiv cs.AI· Hongpeng Cao, Riccardo Curcio, Daniele Ottaviano, Marco Caccamo·· 11 天前AI 评分27
微型神经策略（MNP）实现安全实时机器人控制
Micro Neural Policies for Safe Real-Time Robotic Control
AI 导读
研究者提出微型神经策略（MNP），将进化策略（ES）与统计模型检测（SMC）验证结合用于策略搜索，大幅缩小神经网络规模而不牺牲安全性与鲁棒性。在 Cartpole 和 Quadrotor 任务上完成大规模训练与评估后，MNP 实现零样本 sim-to-real 迁移，内存占用仅 0.5 至 7.5 kB，可在微控制器上部署，实时推理抖动低于 25 ns，芯片空闲时间超过 97%。
正文 · 原文
Abstract:In this paper, we investigate the synthesis of Micro Neural Policies (MNP) to enable safe and robust real-time robotic control on computationally constrained embedded devices. We demonstrate that integrating Evolution Strategy (ES) and Statistical Model Checking (SMC)-based verification for policy search can drastically reduce neural network size without compromising safety and robustness. We conduct a large-scale training and evaluation of MNP on Cartpole and Quadrotor control tasks, varying control frequencies and network architectures. After validating these policies in simulation, we evaluate their deployability through zero-shot transfer to physical systems. Our experiments show that MNP can successfully achieve safe sim-to-real transfer without sacrificing control performance. We then show that the policies' memory footprint, ranging from 0.5 to 7.5 kB, allows deployment on microcontrollers, where they achieve real-time inference latency with under 25 ns of jitter while leaving the chip idle for over 97% of the time for additional workloads. This makes them a highly practical solution for severely resource-constrained robotic systems.
| Comments: | 9 pages |
| Subjects: | Robotics (cs.RO); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.08541 [cs.RO] |
| (or arXiv:2610.08541v1 [cs.RO] for this version) | |
| https://doi.org/10.48550/arXiv.2610.08541 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Hongpeng Cao [view email]
[v1]
Tue, 6 Oct 2026 15:29:28 UTC (1,144 KB)
来源：arXiv cs.AI · arxiv.org