Artificial societies in a community-based approach to ambient intelligence

This paper presents an agent-based approach to the modeling, design and engineering of ambient intelligence systems. The introduced approach balances the management of the complexity related to the enactment of an ambient intelligence scenario between the agents and the structured environment they populate. The dynamics of local actions and interactions of agents situated in this environment originates the overall organized system behaviour. The environment is shaped according to the notion of active coordination artifact, whose seminal definition was given in Computer Supported Cooperative Work literature by Schmidt and Simone and then evolved in other research contexts. A technique to configure and interact with such environments, i.e. the composition of device functionalities according to their high-level features, and services provided, is presented. A scenario is used as an in-depth example, and the architecture of a sample system implementing the same scenario using our reference middleware is discussed. © The Author 2008. Published by Oxford University Press on behalf of The British Computer Society. All rights reserved.