Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Model Learning for Adjusting the Level of Automation in HCPS

Mehrnoush Hajnorouzi, Astrid Rakow, Martin Fränzle

TL;DR

The paper tackles safety in human-centered cyber-physical systems under shared-control by coupling active automata learning of cognitive-model-based human behavior with reactive synthesis to produce correct-by-construction controllers. It operationalizes human behavior as a finite-state abstract model derived from ACT-R simulations, and integrates this HM with a CPS in a timed game to synthesize safety-preserving strategies using Uppaal Tiga. The approach supports iterative refinement: if synthesis or validation fails, the human model or automation variant is revised, linking cognitive modeling directly with formal verification. A driving-case study demonstrates feasibility, showing how learned abstractions and a three-mode supervisory controller can maintain safety while allowing continued human engagement. The work advances principled analysis and design of shared-control HCPS, with implications for robust, explainable automation across safety-critical domains.

Abstract

The steadily increasing level of automation in human-centred systems demands rigorous design methods for analysing and controlling interactions between humans and automated components, especially in safety-critical applications. The variability of human behaviour poses particular challenges for formal verification and synthesis. We present a model-based framework that enables design-time exploration of safe shared-control strategies in human-automation systems. The approach combines active automata learning -- to derive coarse, finite-state abstractions of human behaviour from simulations -- with game-theoretic reactive synthesis to determine whether a controller can guarantee safety when interacting with these models. If no such strategy exists, the framework supports iterative refinement of the human model or adjustment of the automation's controllable actions. A driving case study, integrating automata learning with reactive synthesis in UPPAAL, illustrates the applicability of the framework on a simplified driving scenario and its potential for analysing shared-control strategies in human-centred cyber-physical systems.

Model Learning for Adjusting the Level of Automation in HCPS

TL;DR

The paper tackles safety in human-centered cyber-physical systems under shared-control by coupling active automata learning of cognitive-model-based human behavior with reactive synthesis to produce correct-by-construction controllers. It operationalizes human behavior as a finite-state abstract model derived from ACT-R simulations, and integrates this HM with a CPS in a timed game to synthesize safety-preserving strategies using Uppaal Tiga. The approach supports iterative refinement: if synthesis or validation fails, the human model or automation variant is revised, linking cognitive modeling directly with formal verification. A driving-case study demonstrates feasibility, showing how learned abstractions and a three-mode supervisory controller can maintain safety while allowing continued human engagement. The work advances principled analysis and design of shared-control HCPS, with implications for robust, explainable automation across safety-critical domains.

Abstract

The steadily increasing level of automation in human-centred systems demands rigorous design methods for analysing and controlling interactions between humans and automated components, especially in safety-critical applications. The variability of human behaviour poses particular challenges for formal verification and synthesis. We present a model-based framework that enables design-time exploration of safe shared-control strategies in human-automation systems. The approach combines active automata learning -- to derive coarse, finite-state abstractions of human behaviour from simulations -- with game-theoretic reactive synthesis to determine whether a controller can guarantee safety when interacting with these models. If no such strategy exists, the framework supports iterative refinement of the human model or adjustment of the automation's controllable actions. A driving case study, integrating automata learning with reactive synthesis in UPPAAL, illustrates the applicability of the framework on a simplified driving scenario and its potential for analysing shared-control strategies in human-centred cyber-physical systems.

Paper Structure

This paper contains 31 sections, 6 equations, 4 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Preliminaries
  3. Human--Automation Interaction
  4. Reactive Strategy Synthesis and Safety Games
  5. Levels of Automation and Shared-Control
  6. Iterative Model Learning and Control Synthesis
  7. Generating the Abstract Human Model
  8. Game Graph Construction and Synthesis
  9. Refining the Human Model
  10. Adapting the Automation Design Variant
  11. Preliminary Evaluation through Case Study
  12. HM -- the Driver Model
  13. Scope and rationale.
  14. Learning the Human Behaviour Abstraction
  15. What is identified.
  16. ...and 16 more sections

Figures (4)

  • Figure 1: Overview of the approach: Does the insight into the human and the automation capabilities suffice to implement a safe shared-control HCPS?
  • Figure 2: Learning behaviour abstraction automaton (HM) from simulation of cognitive model.
  • Figure 3: Framework to refine the learned human model HM.
  • Figure 4: Interaction between driver model and driving environment.