Agentic Driving Coach: Robustness and Determinism of Agentic AI-Powered Human-in-the-Loop Cyber-Physical Systems
The paper addresses nondeterminism in agentic AI systems, specifically those used for human-in-the-loop driving coaches. The authors propose using a reactor model of computation (MoC), implemented through Lingua Franca (LF). This approach allows the system to maintain deterministic logic and reliability, even when dealing with the inherent unpredictability of Large Language Models.