Agentic Driving Coach: Robustness and Determinism of Agentic AI-Powered Human-in-the-Loop Cyber-Physical Systems
Deeksha Prahlad, Daniel Fan, Hokeun Kim
Why It Matters
What makes this one worth your time
As AI systems increasingly interact with humans and physical environments, ensuring their reliability and predictability is crucial for safety and effectiveness, particularly in applications like autonomous driving.
The paper proposes a framework to improve determinism in AI-powered human-in-the-loop systems.
Summary
This paper presents a novel approach to enhancing the robustness and determinism of agentic AI in human-in-the-loop cyber-physical systems, specifically through a reactor-model-of-computation framework, with a case study on an agentic driving coach.
Key contributions
- Introduction of a reactor-model-of-computation approach for agentic HITL CPS and a case study demonstrating its application.
Notable insights
- The integration of a reactor-model-of-computation can mitigate the unpredictability of human-AI interactions.
Possible limitations
- The proposed framework may require extensive validation across diverse scenarios to ensure its generalizability and effectiveness.
Abstract
arXiv:2604.11705v1 Announce Type: new Abstract: Foundation models, including large language models (LLMs), are increasingly used for human-in-the-loop (HITL) cyber-physical systems (CPS) because foundation model-based AI agents can potentially interact with both the physical environments and human users. However, the unpredictable behavior of human users and AI agents, in addition to the dynamically changing physical environments, leads to uncontrollable nondeterminism. To address this urgent challenge of enabling agentic AI-powered HITL CPS, we propose a reactor-model-of-computation (MoC)-based approach, realized by the open-source Lingua Franca (LF) framework. We also carry out a concrete case study using the agentic driving coach as an application of HITL CPS. By evaluating the LF-based agentic HITL CPS, we identify practical challenges in reintroducing determinism into such agentic HITL CPS and present pathways to address them.