Formal analysis of safety-critical system simulations

  • Authors:

  • Ayesha Yasmeen;Karen M. Feigh;Gabriel Gelman;Elsa L. Gunter

  • Affiliations:

  • University of Illinois, Urbana, IL;Georgia Institute of Technology, Atlanta, GA;Georgia Institute of Technology, Atlanta, GA;University of Illinois, Urbana, IL

  • Venue:
  • Proceedings of the 2nd International Conference on Application and Theory of Automation in Command and Control Systems
  • Year:
  • 2012

Quantified Score

Hi-index 0.00

Visualization

Abstract

Safety-critical systems are often large and complex. Usually it is not physically or economically feasible to operate these systems under all variant environmental conditions to analyze possible behaviors. Simulating system behaviors under various different environmental conditions and operator guidance patterns provides a cost-effective method of system analysis. In this work we demonstrate how we can formally encode and analyze voluminous simulation traces of safety-critical systems to assess safety and effectiveness requirement conformation. We provide methodology for trace reduction to help obtain tractable yet meaningful formal encoding of the traces. Our methodology is flexible in the sense that one single trace can be analyzed from the point of view of many different properties without having to incur the cost of regenerating the trace. Experimental analysis of a simulation trace can help obtain valuable insights into possible reasons for nonconformance with system requirements. We present our implementation results for traces ensuing from aircrafts attempting to perform Continuous Descent Approach for landing at airport runways. Our work demonstrates that, with the help of faithful abstractions we can obtain valuable insights about simulated traces by the formal verification procedures irrespective of the size of the simulation traces. Formal verification methodology allows for intuitive, expressive yet succinct formulation of system requirements. The combination of simulation trace generation and formal verification provide feedback that may (i) assess the appropriateness of the requirement specifications, (ii) suggest possible infidelity in the simulation modules and (iii) even delineate design error of the original safety-critical system.