Skip to main content
Assured Autonomy Tools Portal
Continual Assurance of Learning-Enabled, Cyber-Physical Systems (LE-CPS)

Tool Support for Assurance Case Development

Abstract

Argument-based\ assurance cases, often represented and organized using graphical\ argument structures, are increasingly being used in practice to provide assurance to stakeholders, e.g., regulatory authorities, that a system is acceptable for its intended use with respect to dependability and safety concerns. In general, comprehensive system-wide assurance arguments aggregate a substantial amount of diverse information, such as the results of safety analysis, requirements analysis, design, verification and other engineering activities. Although a variety of assurance case tools exist, many desirable operations on argument structures such as hierarchical and modular abstraction, argument pattern instantiation, and inclusion/extraction of richly structured information have limited to no automation support. To close this automation gap, over the past four years we have been developing a toolset for assurance case automation, AdvoCATE, at the NASA Ames Research Center. This paper describes how AdvoCATE is being engineered atop formal foundations for assurance case argument structures, to provide unique capabilities for: (a) automated creation and assembly of assurance arguments, (b) integration of formal methods into wider assurance arguments, (c) automated pattern instantiation, (d) hierarchical abstraction, (e) queries and views, and (f) verification of arguments. We (and our colleagues) have used AdvoCATE in real projects for safety assurance, in the context of unmanned aircraft systems.

Year of Publication
2018
Journal
Journal of Automated Software Engineering
Volume
25
Number of Pages
435\textendash499
Date Published
Sep.
DOI
10.1007/s10515-017-0230-5