arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2602.20144eess.SYcs.AIcs.NIcs.SY
收藏

面向可扩展和鲁棒光学系统控制的代理AI

Agentic AI for Scalable and Robust Optical Systems Control

Department of Electrical and Computer Engineering, Duke University(电气与计算机工程系,杜克大学);Duke Quantum Center and Department of Physics, Duke University(杜克量子中心和物理系,杜克大学);Department of Computer Science, Duke University(计算机科学系,杜克大学);NEC Laboratories America(NEC美国实验室);Axiomatic AI(公理AI);Joint Quantum Institute, Department of Physics, and the National Quantum Laboratory (QLab), University of Maryland(联合量子研究所、物理系以及国家量子实验室(QLab),马里兰大学);Research Laboratory of Electronics, Massachusetts Institute of Technology(电子研究实验室,麻省理工学院)
Zehao Wang, Mingzhe Han, Wei Cheng, Yue-Kai Huang, Philip Ji, Denton Wu, Mahdi Safari, Flemming Holtorf, Kenaish AlQubaisi, Norbert M. Linke, Danyang Zhuo, Yira… 展开作者

Zehao Wang, Mingzhe Han, Wei Cheng, Yue-Kai Huang, Philip Ji, Denton Wu, Mahdi Safari, Flemming Holtorf, Kenaish AlQubaisi, Norbert M. Linke, Danyang Zhuo, Yiran Chen, Ting Wang, Dirk Englund, Tingjun Chen

AI总结:

AgentOptics通过代理AI框架实现异构光学系统的可扩展和鲁棒自主控制,展示了高保真任务成功率和多任务协调能力。

摘要:

我们提出了AgentOptics,一种基于模型上下文协议(MCP)的代理AI框架,用于高保真的自主光学系统控制。AgentOptics通过结构化的工具抽象层解释自然语言任务,并在异构光学设备上执行符合协议的行动。我们实现了64个标准化的MCP工具,覆盖8种代表性的光学设备,并构建了一个410任务基准测试,以评估请求理解、角色感知响应、多步协调、对语言变异的鲁棒性以及错误处理。我们评估了两种部署配置——商业在线LLM和本地托管的开源LLM,并将其与基于LLM的代码生成基线进行比较。AgentOptics实现了87.7%至99.0%的平均任务成功率,显著优于代码生成方法,后者最高仅达到50%的成功率。我们还通过五个案例研究展示了更广泛的应用,从设备级控制扩展到系统编排、监控和闭环优化。这些包括DWDM链路配置和相干400 GbE及模拟光纤(ARoF)通道的协同监控;自主表征和偏置优化宽带ARoF链路承载5G前传流量;多跨度通道配置与发射功率优化;闭环光纤偏振稳定化;以及基于分布式声学传感(DAS)的光纤监控与LLM辅助事件检测。这些结果确立了AgentOptics作为自主控制和编排异构光学系统可扩展和鲁棒范式的基础。

英文摘要:

We present AgentOptics, an agentic AI framework for high-fidelity, autonomous optical system control built on the Model Context Protocol (MCP). AgentOptics interprets natural language tasks and executes protocol-compliant actions on heterogeneous optical devices through a structured tool abstraction layer. We implement 64 standardized MCP tools across 8 representative optical devices and construct a 410-task benchmark to evaluate request understanding, role-aware responses, multi-step coordination, robustness to linguistic variation, and error handling. We assess two deployment configurations--commercial online LLMs and locally hosted open-source LLMs--and compare them with LLM-based code generation baselines. AgentOptics achieves 87.7%--99.0% average task success rates, significantly outperforming code-generation approaches, which reach up to 50% success. We further demonstrate broader applicability through five case studies extending beyond device-level control to system orchestration, monitoring, and closed-loop optimization. These include DWDM link provisioning and coordinated monitoring of coherent 400 GbE and analog radio-over-fiber (ARoF) channels; autonomous characterization and bias optimization of a wideband ARoF link carrying 5G fronthaul traffic; multi-span channel provisioning with launch power optimization; closed-loop fiber polarization stabilization; and distributed acoustic sensing (DAS)-based fiber monitoring with LLM-assisted event detection. These results establish AgentOptics as a scalable, robust paradigm for autonomous control and orchestration of heterogeneous optical systems.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑