Presentation Information
[10a-N304-6]Proposal and Prototyping of an Edge-AI Measurement Device with Multi-Standard Interfaces for the Era of Self-Driving Laboratories
〇Masayuki Fukuoka1 (1.Sankaku Mfg.)
Keywords:
Self-Driving Laboratory,Materials Informatics
As self-driving laboratories (SDLs) and generative AI now operate physical instruments directly, measurement devices must offer not only a human GUI but also standardized interfaces for IT systems, robotics, and AI agents. Most existing SDLs absorb this on the platform side, retrofitting each instrument via custom adapters or vendor-assisted API changes, making integration cost a barrier. We instead propose a principle in which the device itself embeds multiple standard interfaces and an edge-AI layer, becoming a reusable node that plugs into any SDL, data platform, or AI agent with no extra work while structuring data at its source. We define connectivity for four user classes—human (GUI/tabular data), IT (REST/OpenAPI), robotics (OPC UA), and generative AI (MCP)—and implement them in parallel on one edge device. We report LabCam, a sample-imaging prototype, confirming identical imaging and data acquisition via all four paths, and discuss design guidelines for the era of Physical AI.