A Graph-Model-Based Testing Method Compared with the Classification Tree Method for Test Case Generation

Abstract

In automotive industry, mechanic control units are more and more replaced by electronic devices that are often aggregated in electronic control units (ECU). Systematic testing is one of the preferred industrial validation methods to ensure functionality of those ECUs. It is imperative to reduce the costs and improve the effectiveness of testing by automating the testing process. This paper introduces a model-based testing method using event sequence graphs and compares this approach with the classification tree method which is popular in automotive industry. A case study applies both methods to the formal specification of an adaptive cruise control unit for generation and selection of test cases. To enable a meaningful comparison, test costs and number of faults revealed by both methods will be compared.