FURI | Spring 2026

Deterministic Safety of AI Agent-Embodied Robots Using a Model-Based Design Approach and Lingua Franca Open-source Framework for Cyber-Physical Systems.

This research investigates how deterministic and time-predictable behavior can be enforced in artificial intelligence (AI)-powered robotic agents operating within safety-critical cyber-physical systems (CPS). Nondeterministic AI outputs combined with unpredictable human behavior introduce dangerous timing variability in human-in-the-loop (HITL) systems, posing serious risks in real-world environments. Using the Lingua Franca (LF) coordination framework with a large language model (LLM)-based robotic agent in a MuJoCo physics simulation, this approach aims to identify measurable improvements in safety response reliability relative to an uncoordinated baseline. Future work will explore deploying this framework on physical robotic hardware in real-world human-robot interaction scenarios.