Abstract
The satisfaction of Non-Functional Requirements (NFRs) has a clear impact in the success of Ambient Intelligence (AmI) Systems. However, existing methods and techniques to engineer these systems lack support in their specific and systematic treatment. In this paper, a Model-Driven Engineering (MDE) method for the design of AmI systems is introduced to deal with NFRs paying special attention to those particularly related to the features of these kind of systems, namely, unobtrusiveness, context-awareness, and self-adaptativeness. The aim is to enable the semi-automatic derivation of quality software design models from an NFRs specification of the system, with special focus on agent-oriented architectures. To this end, Computation Independent Models (CIM) and transformation techniques between them are defined, as well as an evaluation procedure to be applied, which can help developers to make successful solution decisions satisficing the overall quality properties required for AmI systems, and also specific ones for each particular system. The proposal is illustrated by means of a case study of an Ambient Assisted Living (AAL) system for elderly people.