A TOOL FOR THE DESIGN AND ANALYSIS OF SOFTWARE SAFETY-CRITICAL SYSTEMS

摘要

This paper describes a tool for the design and analysis of military systems with the objective of enhancing their safety-critical success. The tool uses Artificial Intelligence (AI) techniques to identify and link undesired system events to their mitigating requirements. An effective implementation of this tool requires all system events to be described in the form of linguistic inference flow charts, such as fault trees. Strategies to automatically process linguistic information include fuzzy logic, neural networks and expert rules. The DOORS project development and management software provides the platform upon which this tool is built. Also, where appropriate, subsystems are designed with the use of adaptive models, with the objective of enforcing safety requirements at subsystem interfaces. This paper highlights the development of this tool in relation to legacy systems, such as the H-60, and potentially to future systems such as Joint Strike Fighter (JSF).