SIMULTATE: A Toolset for Fault Injection and Mutation Testing of Simulink Models

Abstract

The advantages of fault injection techniques and related methodologies like mutation testing have been gaining in attention also from industry, as is evident from the advent of standards like ISO 26262 that suggest to use corresponding approaches for verifying an automotive system's safety aspects. Aside a well-established theoretical background, the availability of tools is a key issue in order to leverage fault injection for the development of industrial, possibly safety-critical applications, e.g., in an automotive context. We propose the corresponding open source toolset SIMULTATE for injecting faults and performing mutation testing for Simulink models. For complementing the provided mutation/fault injection operators, it allows a user to define her own ones within Matlab and further provides a Python interface for easily deriving mutants where she can also focus the scope to desired model parts only. Controlling the activation of individual faults in a derived model, a designer can conveniently conduct mutation tests via a corresponding Python application.