Model-Driven Security Test Case Generation Using Threat Modeling and Automata Learning

Automotive systems are not only becoming more open through developments like advanced driving assistance functions, autonomous driving, vehicle-to-everything communication and software-defined vehicle functionality, but also more complex. At the same time, technology from standard IT systems become frequently adopted in this setting. These developments have two negative effects on correctness and security: the rising complexity adds potential flaws and vulnerabilities while the increased openness expands attack surfaces and entry points for adversaries. To provide more secure systems, the amount of verifying system security through testing has to be significantly increased, which is also a requirement by international regulation and standards. Due to long supply chains and non-disclosure policies, verification methods often have to operate in a black box setting. This thesis strives therefore towards finding more efficient methods of automating test case generation in both white and black box scenarios. The focus lies on communication protocols used in vehicular systems. The main approaches used are model-based methods. We provide a practical method to automatically obtain behavioral models in the form of state machines of communication protocol implementations in real-world settings using automata learning. We also provide a means to automatically check these implementation models for their compliance with a specification (e.g., from a standard). We furthermore present a technique to automatically derive test-cases to point out found deviations on the actual system.We also present a method to create abstract cybersecurity test case specifications from semi-formal threat models using attack trees.