Design and Evaluation of an Executable Test Case Script Generator Framework Based on Natural Language Specifications in Automotive E-Mobility ECU Testing
CovenanteeTechnische Universität München
Document typeMaster thesis
Rights accessRestricted access - author's decision
All rights reserved. This work is protected by the corresponding intellectual and industrial property rights. Without prejudice to any existing legal exemptions, reproduction, distribution, public communication or transformation of this work are prohibited without permission of the copyright holder
Testing is crucial to develop reliable automotive software successfully and represents a vital part of the software’s life cycle as it verifies the correct behaviour of the implemented functionality. In order to identify bugs in the software, test cases must be designed and developed to be executed in a specific testing platform such as hardware-in-the-loop or test bench. Obtained testing results can improve the software code and assure that the implemented functionalities behave as expected. This thesis aims to improve software testing efficiency in terms of time savings and quality of test cases by automating part of the process for hardware-in-the-loop and test bench testing platforms of powertrain e- mobility automotive electronic control unit (ECU) software and its integration with other components such as the battery or the electric machine. For each software version of the automotive ECU, different tests must be conducted in different testing platforms to ensure that the software behaves as expected and that the interaction with the whole system fulfils the needed requirements. This process is highly automatable since executable test case scripts can perform these tests without any other human interaction. However, test scripts are explicitly programmed for a particular environment and cannot be reused in other platforms. In addition, executable test scripts are developed manually, which is human error-prone and labour-costly. Therefore, to improve current testing methods, the peculiarities of each testing environment and the overall testing workflow are analyzed to identify possible improvement opportunities later and present a viable solution that improves the current efficiency. This thesis presents an approach to increase the testing efficiency of powertrain e- mobility ECUs, which automatically generates executable test case scripts based on test case specifications in natural language for hardware-in-the-loop and test bench testing environments. This generator creates the XML code using the file format of the program ECU-Test later to allow the automatic execution in the testing platform. The central concept for the automatic generation of test case scripts is structured in two stages with an intermediate generic format in between. The generator is template-based and interprets the natural language by the use of keywords. In order to support the developed methodology, a generator application is developed as a proof of concept.
DegreeMÀSTER UNIVERSITARI EN ENGINYERIA INDUSTRIAL (Pla 2014)