PUDA: An AI-Native Hardware Harness for Self-Driving Laboratories

2026-07-29

AI Summary

This work addresses the limitations of conventional self-driving laboratories, which rely on human-centric graphical interfaces that hinder efficient and reliable control of experimental hardware by AI agents. To overcome this, the authors propose the first AI-native, headless hardware execution environment, leveraging a command-line interface, a structured JSON protocol, and a distributed messaging system to unify device discovery, command routing, and data artifact management. The framework introduces a hardware abstraction layer specifically designed for AI agents, decoupling scientific orchestration from physical execution. By organizing all operations around run IDs and timestamps, it establishes an end-to-end traceable data structure that guarantees determinism, atomicity, and auditability. This infrastructure provides a standardized, scalable, and verifiable foundation for hardware interaction in physical AI systems.

Abstract

Physical Unified Device Architecture (PUDA) is an AI-native hardware harness for self-driving laboratories (SDLs). Rather than building a human-centered graphical user interface (GUI) orchestration layer, PUDA creates a command-line runtime environment that lets agents observe, orient, decide, and act over experiments while hardware execution remains deterministic, atomic, and auditable. Headless by design, devices appear through discoverable command-line interfaces, JSON protocols are routed through a distributed messaging system, and command responses, data products, and reports are preserved as structured records. PUDA organizes protocols, runs, samples, measurements, and command logs into an AI-native data structure linked by run identifiers and timestamps, preserving provenance from submitted protocol through hardware response to resulting data products. PUDA separates scientific orchestration from physical operation and data telemetry: agents choose experiments, while PUDA executes validated commands and captures provenance-linked state, responses, and data. The contribution is not another optimizer, orchestrator, or recipe language. It is a practical execution and data environment for agentic SDLs; the broader physical AI implication is that PUDA provides an AI-native hardware harness for AI systems to interact with physical tools.

Problem

• self-driving laboratories
• AI-native hardware
• provenance tracking
• deterministic execution
• physical AI

Innovation

• AI-native hardware
• self-driving laboratories
• command-line runtime
• provenance tracking
• deterministic execution