Clone Detection in Matlab Simulink Models

Abstract

A growing amount of embedded software is created by automated code generation from models. As each development project requires a given level of software quality, it is essential for successful projects that a projects quality is monitored and assessed continuously during the development life cycle. While quality assessment tools and processes for conventional software engineering are widely available, quality assessment for models is a fairly young research topic. Clone detection in models is one method to assess and improve a model’s maintainability. Matlab Simulink is one of the leading tools for model based software development in the automotive industry. In this thesis a novel clone detection algorithm is proposed. The algorithm was developed by analyzing the structure and characteristics of graph based models in general and Simulink models in particular. In contrary to existing clone detection algorithms the new algorithm follows a top-down approach for identifying potential cloned parts of a model. Further the proposed solution is prototypically implemented in a clone detection tool for Matlab Simulink models.