Finding Dependency, Test Sequences and Test Cases for Simulink/Stateflow Models

Abstract

The Simulink/Stateflow (SL/SF) acquiring from Mathworks is fitting the de facto standard in industry for model based development especially for embedded control systems. Many industrial tools are available in the market for test case generation from SL/SF designs; though, we have observed that these tools do not accomplish satisfactory coverage in cases when designs involve non-linear blocks and Stateflow blocks transpire deeper inside the Simulink blocks. For this purpose, we have proposed a methodology that generates the test sequences and test cases from the Simulink/Stateflow model. In our approach, first, we have developed a SL/SF model using MATLAB tool which generates mdl(model description language) file. Next, we convert that mdl file into xml file and then the xml file and mdl file path are passed as an inputs to our proposed methodology to generate Simulink/Stateflow dependency graph(SSDG); Now using the SSDG, we generate test sequences by applying depth first search approach(DFS). Next, for each test sequence, we generate a set of test cases and finally we prioritize those test cases using information flow(IF)value. Sl/SF model allows modelling the systems, simulating and analyzing dynamic systems. The resultant Simulink/Stateflow models consist of large numbers of blocks and states likes more than ten thousand blocks. Hence, to certify the quality of such control system models, automated static analyses and slicing methods are necessary to deal with this complexity. Hence, these approaches help in debugging the model, understanding the behavior of models,identifying faults,if occurs. In this thesis, we present an approach for computing intra-dependencies between blocks by the concept of slicing approach and we represent the result using dependence graphs. With the help of slicing approach, the complexity of a system model can be compact to a specified point of interest(Slicing Criterion) concern by removing unrelated blocks in model system.