Agent-based intelligent middleware for user-centric services in ubiquitous computing environments

Abstract

With the emergence of the ubiquitous computing era, it has become increasingly important for a middleware that fully exploits Human Computer Interaction (HCI)-friendly characteristics in order to support user-centric services. Many studies about HCIfriendly middleware designed to support user-centric services have been performed. However, most of the existing research focuses on functional improvements to middleware that concentrates on a specific application rather than on general user-centric services. In this paper, we propose an agent-based intelligent middleware that provides usercentric services in ubiquitous computing environments. We introduce various HCIfriendly middleware projects that support user-centric services, and propose a design for a middleware architecture that effectively supports several functions, such as context awareness, context management, pattern inference based on user behavior, and dynamic agent generation. We also suggest service scenarios that illustrate the adaptive User Interface (UI), spontaneous UI migration for collaboration between users, and UI migration according to the changes of the user's task environments in smart spaces. We implemented a prototype of our middleware architecture, called Agent-based Intelligent Middleware (AIM), based on our scenarios, in order to show the benefits by verifying the feasibility of user-centric services with minimal user intervention in ubiquitous computing environments.