![[2602.20144] Agentic AI for Scalable and Robust Optical Systems Control](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.20144] Agentic AI for Scalable and Robust Optical Systems Control
Summary
The paper presents AgentOptics, an AI framework for autonomous control of optical systems, achieving high task success rates and demonstrating broader applications in system orchestration.
Why It Matters
This research is significant as it showcases the potential of agentic AI in enhancing the efficiency and reliability of optical systems control. By outperforming traditional code generation methods, it opens new avenues for automation in telecommunications and networking, which are crucial for modern infrastructure.
Key Takeaways
- AgentOptics framework interprets natural language tasks for optical systems.
- Achieves 87.7% to 99.0% task success rates, outperforming code generation methods.
- Demonstrates applicability in various case studies beyond device control.
- Utilizes a structured tool abstraction layer for heterogeneous optical devices.
- Highlights the scalability and robustness of AI in optical systems orchestration.
Electrical Engineering and Systems Science > Systems and Control arXiv:2602.20144 (eess) [Submitted on 23 Feb 2026] Title:Agentic AI for Scalable and Robust Optical Systems Control Authors: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 View a PDF of the paper titled Agentic AI for Scalable and Robust Optical Systems Control, by Zehao Wang and 14 other authors View PDF HTML (experimental) Abstract: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 studie...