Agent-Based Group Modelling for Ambient Intelligence

Ambient intelligence allows physical environments to become sensitive and responsive to the presence of people and objects. An environment endowed with ambient intelligence is able to analyse its contexts, adapt itself to the presence of people and objects residing in it, learn from their behaviour and recognise and express emotion. Ambient intelligence is realised via devices which blend into the background, while supporting social interaction and improving people’s experience within the physical space (e.g., by increasing safety or comfort). Often physical spaces must be shared by various people: adapting devices ’ responses and behaviour to simultaneously suit a group of people is an important and not much explored issue. To complicate matters further, group membership may change continuously. In this paper, we propose an approach based on group adaptation and software agents, to manage shared devices in ambient intelligence solutions. We present a proof-of-concept implementation embedding our approach, which allows engineers to design and experiment with distinct ways of managing shared devices – software agents are associated with devices and people, and interact with each other to agree on how shared devices should change their behaviours in view of the people in their radius of action.