The thesis aims at designing new interaction paradigms for smart environments with a particular focus on environment-to-user communication. Messages generated by a smart environment, either by means of inner processes (e.g., detection of some internal status like a door opening) or by bridging external services (e.g., e-mail, chat, etc.) have to be routed to the right recipient, on the most suitable device, depending on the current user location, activity and surrounding context (e.g., noisy vs quiet, dim light vs bright illumination, etc.).
The main focus of the work involved 2 main aspects:
- the creation of a suitable modeling infrastructure for end-user device capabilities, user activities, for the current context in a home setting and for the type of messages generated / delivered by the smart home;
- the creation of a software infrastructure able to deliver messages to selected end-user devices, and to capture context information from the user surroundings (might / should exploit existing systems, e.g., Dog), accounting for user preferences and possibly negotiating with the user the most suitable interaction pattern.
To fulfill the previous goals several simplifying assumptions were applied, including:
- users' location is assumed to be known;
- users' identity is assumed to be known;
- end-user devices may only be in a pre-defined set (e.g., mobile phone, smart watch, audio diffusion, tv, pc screen, in-home display).