Vehicles have come a long way from being purely mechanical systems to systems that consist of an internal network of more than 100 microcontrollers and systems that communicate with external entities, such as other vehicles, road infrastructure, the manufacturer’s cloud and external applications. This combination of resource constraints, safety-criticality, large attack surface and the fact that millions of people own and use them each day, makes securing vehicles particularly challenging as security practices and methods need to be tailored to meet these requirements.
This thesis investigates how security demands should be structured to ease discussions and collaboration between the involved parties and how requirements engineering can be accelerated by introducing generic security requirements. Practitioners are also assisted in choosing appropriate techniques for securing vehicles by identifying and categorising security and resilience techniques suitable for automotive systems. Furthermore, three specific mechanisms for securing automotive systems and providing resilience are designed and evaluated.
The first part focuses on cyber security requirements and the identification of suitable techniques based on three different approaches, namely (i) providing a mapping to security levels based on a review of existing security standards and recommendations; (ii) proposing a taxonomy for resilience techniques based on a literature review; and (iii) combining security and resilience techniques to protect automotive assets that have been subject to attacks.
The second part presents the design and evaluation of three techniques. First, an extension for an existing freshness mechanism to protect the in-vehicle communication against replay attacks is presented and evaluated. Second, a trust model for Vehicle-to-Vehicle communication is developed with respect to cyber resilience to allow a vehicle to include trust in neighbouring vehicles in its decision-making processes. Third, a framework is presented that enables vehicle manufacturers to protect their fleet by detecting anomalies and security attacks using vehicle trust and the available data in the cloud.