The everyday schedule and habits of each person result in a sequence of actions that can be observed, which can also be referred to as their behaviour. An in-depth context analysis is performed on a smart home environment to create a context-aware solution for recognizing patterns in the behaviour of the users and the system. These patterns are made available for access control, which makes it possible for the access control mechanism (ACM) to include the knowledge of the usual behaviour found in the system during the access control decision. The central questions of this thesis are how behaviour can be modeled as behaviour patterns in a smart home environment, how the modeled behaviour can be used in access control, and how long the learning time of these behaviour patterns is. The behaviour engine is introduced as a smart home component responsible for determining behaviour and an extensive evaluation is performed against it to confirm the results of this thesis. The evaluations are partly based on CosyHome [25], which is a big dataset created in a case study as part of this work.
