Agent-Oriented Context Collection and Customised Service Delivery

A key issue in pervasive computing is the trade-off between personalised services (e.g. according to context) and volunteering information (i.e. the context itself) in order to secure those services. In this paper, we propose a novel agent-based architecture to facilitate collection of contextual information and the delivery of personalised services, where both collection and delivery processes are sensitive to the business model underpinning the application domain. We explicate this architecture through a ´virtual clipboard´ scenario, in which relevant content is provided for the highlights of a museum visit. The content is delivered according to contextual information, which includes an estimation of each visitor´s emotive state. This is computed from physiological sensor readings obtained during human-computer interactions. The architecture thus provides a platform to investigate the extent to which a user is prepared to reveal information and to forgo certain aspects of privacy in order to leverage the benefits of added-value services