In this paper, we describe our work in developing an autonomic system that supports the composition of ubiquitous applications at run-time. The applications are composed and adapted accordingly to user preferences and user-provided criteria. We have designed a proof-of-concept prototype of the system and an example multimedia application. The application is a multimedia player which users can control on a large screen using the mobile phone's UI as the controller. We present a user evaluation of the prototype's feasibility and also determine what feedback and control mechanisms are required by the end-users. We report initial analysis of how user satisfaction and comfort level are affected by the autonomy of the system.

User profiles are increasingly used for sharing standard information about users among context-aware agents. User profiles allow agents to offer users personalized content and services. However, the entities and contextual information used by these agents must have same meaning in order to share a common understanding about user related personal information, context and preferences. The contribution of this paper is to present a general user metadata model which is integrated within a generic metadata model (CAM Meta-model) that covers altogether information about content, services, physical and technical environment. This new user profile meta-model has been designed with a view of using it in conjunction with content and service recommender systems. It brings new opportunities to reason over user context data with the main purpose of increasing user experience in ubiquitous environments and satisfying their desires depending on the circumstances.

No abstract