An Ontology Based Approach to Interaction Ambient Design Augusto Celentano, Andrea Okroglic, Fabio Pittarello Universit` a Ca’ Foscari di Venezia, Dipartimento di Informatica Via Torino 155, 30172 Mestre (VE), Italy {auce,aokrogli,pitt}@dsi.unive.it Abstract—We discuss the design of interaction ambients, which are ambients populated with fixed and mobile services, accessible through distributed interfaces. To make the interaction effective, activities and services should be designed as a function of the ambient, giving the user a coherent and clear context. A design approach is discussed, based on ontologies describing ambients, activities and services. I. I NTRODUCTION This paper discusses a methodology for designing inter- active applications in complex physical environments, where human activity and interaction need to be adapted to the environment according to its functions. We shall denote the set of functions, roles and purpose of a physical environment with the expression ambient semantics; an environment in which services are offered through communication and information processing access points will be called interaction ambient. In general terms, an application can be viewed as a set of cooperating services executing tasks and exchanging data, providing computation and information support, as well as specific results, to humans in a coordinated plan of actions. We use the term service in a generic sense, without subsuming any specific service architecture. Services execute (sub-)tasks, exchange information and objects, process information content and presentation, etc., according to a plan. Services may also be executed by humans: for example, an information service can be offered, in different contexts, by an automatic answering system, an interactive kiosk or a human guide. The methodology represents a transition from context- awareness to ambient-awareness. Context-aware applications are able to adapt their behavior to a number of parameters that may belong to unrelated classes, such as location, time, device, user profile, etc., by selecting proper services among a set of alternatives, or by processing the information content to suit the constraints of the equipment. Context-aware applications are usually associated to mobile and ubiquitous devices: due to their limited processing and presentation features, adaptation helps the user to get information through these devices in less time and with less effort [1]–[4]. Shifting from context-awareness to ambient-awareness, adaptation depends not only on local properties of the user and her/his context, but also on the ambient role, i.e., on its meaning as a private or public place, having a defined func- tion, a physical structure, and a space organization strongly depending on its function (such as a school, a hospital, an airport). An ambient-aware application appears as a set of pervasive services distributed in a physical environment, accessible from local appliances, or within a wide area, or through mobile devices, delivering information and tangible objects, with which users interact in ways tailored to their context and to the properties, role and goal of the ambient. Adaptation relates to the identification of the relevant prop- erties of the environment navigated by the user. In most cases they are high-level properties (e.g., functional or social fea- tures) that can’t automatically be inferred from measurements derived from a sensors’ network. The combined geometric and semantic description of real environments is a promising research fields that can considerably help to improve the design of interactive distributed applications [5], [6]. II. ELEMENTS OF THE APPROACH The study of the relations between the environment, its embedded services and the user activities evolving in it is a pre-requisite for optimizing the overall interactive system and obtaining the best results in terms of user satisfaction. We consider the different constraints that can affect the original state of the interactive system, discussing which are the potentialities and the optimization constraints in each situation. A. Definitions We use two basic definitions for the environment: ambient and location. The term ambient is referred to a spatial area organized in locations and populated by artifacts (e.g., urban infrastructures, buildings, embedded objects, etc.) and natural elements (e.g., rivers, plants, etc.); the term location identifies a partition in the space delimited by different types of objects (e.g., walls, floor, doors), morphologically meaningful for the user and suitable for the development of different classes of user activities. The definition for location is derived from a previous research activity targeted at identifying the structure of three dimensional environments belonging to the wide domain of mixed reality [7]. The term interaction ambient is referred to an environment characterized by the availability of embedded communication technologies that may be accessed by a visitor in order to get information and services. The term service denotes a work executed by a provider for a consumer, producing some kind