Detection of Deployment Operational Deviations for Safety and Security of AI-Enabled Human-Centric Cyber Physical Systems

TL;DR

This work tackles safety and security in AI-enabled human-centric CPS by focusing on deployment operational deviations that arise under uncertain human interactions. It proposes a personalized, data-driven framework that includes a detection model and thresholds, plus an image-based method to detect missed meal announcements in an artificial pancreas. A key contribution is the time-series image encoding of glucose–insulin dynamics into recurrence-plot images and a CNN-based classifier operating on those images, underpinned by a sensitivity-relations framework with a matrix $SR$ and insulin-sensitivity estimates $SI_i$ derived from a minimal model. Experimental results on outpatient Type 1 diabetes data show promising but variable performance across individuals, underscoring the benefits of personalization and signaling the need for improved explainability and generalization to other AI-enabled CPS such as autonomous vehicles.

Abstract

In recent years, Human-centric cyber-physical systems have increasingly involved artificial intelligence to enable knowledge extraction from sensor-collected data. Examples include medical monitoring and control systems, as well as autonomous cars. Such systems are intended to operate according to the protocols and guidelines for regular system operations. However, in many scenarios, such as closed-loop blood glucose control for Type 1 diabetics, self-driving cars, and monitoring systems for stroke diagnosis. The operations of such AI-enabled human-centric applications can expose them to cases for which their operational mode may be uncertain, for instance, resulting from the interactions with a human with the system. Such cases, in which the system is in uncertain conditions, can violate the system's safety and security requirements. This paper will discuss operational deviations that can lead these systems to operate in unknown conditions. We will then create a framework to evaluate different strategies for ensuring the safety and security of AI-enabled human-centric cyber-physical systems in operation deployment. Then, as an example, we show a personalized image-based novel technique for detecting the non-announcement of meals in closed-loop blood glucose control for Type 1 diabetics.

Figures (5)

• Figure 1: System model
• Figure 2: Detection Architecture
• Figure 3: Workflow
• Figure 4: Snack Data
• Figure 5: Accuracy and F1-Score