Autonomous Vehicles (AVs) are complex safety-critical systems that operate in uncertain and dynamic environments. During runtime, the environmental uncertainties and random component failures might result in hazardous events, and sometimes even to an accident, if left undetected. AVs must interact with different elements from their surrounding environment- like pedestrians and traffic scenarios and provide safe functional behaviour. One way to achieve safe behaviour even in the presence of adverse situations or abrupt component failures is through self-adaptation. Self-adaptation helps an AV to automatically adapt and modify itself in response to changes in its surrounding environment. In [1], Salehie and Tahvildari propose a set of six fundamental questions which must be addressed to implement self-adaptation characteristics: (1) When to adapt? (2) Why do we have to adapt? (3) Where do we need to implement change? (4) What kind of change is needed? (5) Who has to perform the adaptation? and (6) How is the adaptation performed? We only address the question: (what) kind of adaptation should be carried out on what and how many different levels. Additionally, we also aim to identify the necessary requirements for implementing system adaptation.