ISBN 978-3-8439-1819-0

978-3-8439-1819-0, Reihe Informatik

Philipp Luchscheider
Strategic Black-Box Testing of Automotive Safety Electronic Control Units

204 Seiten, Dissertation Universität Erlangen-Nürnberg (2014), Softcover, A5

Zusammenfassung / Abstract

Innovations in modern premium vehicles are nowadays made by new functions for driver assistance and active passenger safety more than ever before. Features like the adaptive cruise control and the autonomous emergency brake as well as enhanced passive safety are reducing the risk of fatal injury in the event of a car accident. Thus, the realized functions are getting more complex and interoperable, the developed Electronic Control Units (ECUs) become more powerful and interconnected. The newly developed functions require new technologies of vehicle context perception and increase the amount of exchanged data inside the vehicle. This in contrast may increase the damage to the vehicle and the passengers in the event of a system malfunction, when driving situations are incorrectly interpreted. In order to reduce the chance of erroneous behavior, the efforts for designing and testing these systems increase. Raising test efficiency is needed to be able to develop ECUs safely and reliable with having short development cycles.

One main innovation in this thesis is using the combined abstract views of the inside of the system and the black-box test cases during analysis. This linkage enables profiling test suits with the knowledge about aspects of development, like restraints by the iterative development and safety issues. For example, test cases can be detected without being executed, that would give invalid feedback in early stages of development. Furthermore, after an update or noticing peculiarities of parts of the system, test cases can be found to further investigate those parts. Two different application examples were set up in the context of the research project between the University Erlangen-Nuremberg and the German car manufacturer Audi. Potential savings in the development cycles as well as parts of the system with higher need for testing could be identified right at the beginning of the development. The results demonstrate the applicability and the flexibility of the developed approach by having two system with different levels of specification and unequal test environments.