AB-HCI: an interface multi-agent system to support human-centred computing

Currently, new types of information are available to applications. This incoming stream of data can be processed to improve user experience. User experience can be enhanced to present contents to users in a better manner, or to provide software artefacts to improve their interaction with the environment. The wide range of interaction design capabilities offered by this data requires methods and run-time architectures able to cope with the integration of the incoming data from the interaction environment into the applications. Human-centred computing takes into account what the state of user interaction is and how this data can be used to actually make it useful. A multi-agent system architecture is described. This architecture is able to tackle the arrival of context data by means of a set of sensors and interface agents that filter and process the context-of-use data to produce new presentations. The context of use provides information about the user, the platform, the physical environment and the current task the user is performing. These new presentations produced are adapted according to the context-of-use information gathered by applying a mixed-initiative approach to overcome some of the usual flaws found in full-adaptive applications.