C. Stephanidis (Ed.): Universal Access in HCI, Part II, HCII 2007, LNCS 4555, pp. 1052–1061, 2007. © Springer-Verlag Berlin Heidelberg 2007 Personal Companion: Personalized User Interface for U-Service Discovery, Selection and Interaction * Hyoseok Yoon, Hyejin Kim, and Woontack Woo GIST U-VR Lab., Gwangju 500-712, South Korea {hyoon,hjinkim,wwoo}@gist.ac.kr Abstract. In this paper, we propose a mobile user interface named personal companion which enables selection and interaction of u-services based on context of user. Personal companion selects u-services from a list of discovered services, supports camera-based selection with embedded marker and personalizes UI of the selected service in ubiquitous computing environment. In order to verify its usefulness, we implemented personal companion on PDA and UMPC platform and deployed into smart home testbed for selecting and interacting with u-services. The proposed personal companion is expected to play a vital role in ubiquitous computing environment by bridging users and u- services. Keywords: mobile user interface, camera-based interaction, personalization, embedded marker. 1 Introduction Recent interest and increased awareness for future computing paradigms such as ubiquitous computing and context-aware computing are getting more attention from both researchers and end-users alike. In these future computing environments, users are served by pervasive and ubiquitous intelligent services for highly personalized and easy-to-use services. Ultimately, the goal is to realize the vision of Mark Wesier who envisioned technologies which disappear into everyday life of users [1]. However, the traditional desktop-based interface and interaction methods have limitations in getting and using the services in the ubiquitous computing environment (u-service, hereafter). This limitation arises, because u-service is characterized by properties such as ubiquity and context-awareness. Especially ubiquity of u-service builds upon the assumption of mobility of users, and context-awareness requires acquiring and analyzing information of users and the environment. Moreover, all these functionality should be available anywhere and anytime in a form of compact- sized computing resource. A viable approach is introduced with recent advancement * This research is supported by the UCN Project, the MIC 21 st Century Frontier R&D Program in Korea.