sOc-EUSAI'2005 conference

Regular session - Context-aware systems
P1-1 GENIO : An Ambient Intelligence application in home automation and entertainment environment
Garate, Alfonso - Herrasti, Nati - Lopez, Antonio
As formated for the printed proceedings - 10.ps - 10.pdf - pages 241-246
As delivered by the authors - 10_pdf_file.pdf
Abstract :
As one of the main aims of "Ambient Intelligence" is natural human interaction with the environment and one the most suitable is the home environment. Fagor has been working several years to develop a power-line network where all its household appliances, security sensors and actuators, heating systems and antiintrusion systems are connected and managed by a central controller named Maior-Domo. As a result of GENIO project, the user can dialog with his home and asks the services and functionalities he wants by talking as he was talking to a friend. The controller Maior-Domo has a human representation that the user can see and can interact with. When the user talks the Maior-Domo extracts the different commands from those vocal orders and controls the home devices. These orders are not specific commands that the user has to learn but natural speaking language without any need of learning. In the same way, any event or information from any device of the network is transmitted to the user by voice. In order to achieve a demonstrator of this Ambient Intelligence application, a real kitchen and sitting room have been built where the users can command the home talking naturally (in Spanish). Possible actions are: reading e-mails, programming the washing machine, checking the goods in the fridge, creating the shopping list, doing shopping with a PDA in the supermarket, activating the dishwasher, being guided on how to prepare a recipe for the oven checking if there are the needed goods to do it, listening some music stored at home, watching some photos, watching some selected video and so on. Every user has a wireless microphone in his/her shirt.s pocket. This microphone captures his/her voice and all the sounds around him/her and sends them to a developed board which filters the voice frequency range from other sounds. From here the voice recognition system "understands" the pronounced sentence and process it. A quite extended number of sentences, called grammar, make up the possible dialogue between the person and the whole system. The user can address the whole system in different ways using a lot of expressions, talking naturally and spontaneously and dialoguing to the home. The defined grammar is so extended that almost the total speaker independence has been achieved.


P1-2 Basic components of a Face-to-Face interaction with a Conversational Agent : Mutual Attention and deixis
Raidt, Stephan - Bailly, Gerard - Elisei, Frederic
As formated for the printed proceedings - 65.ps - 65.pdf - pages 247-252
As delivered by the authors - 65_pdf_file.pdf
Abstract :
We present a series of experiments that involve a face-to-face interaction between an embodied conversational agent (ECA) and a human interlocutor. The main challenge is to provide the interlocutor with implicit and explicit signs of mutual interest and attention and of the awareness of environmental conditions in which the interaction takes place. A video realistic talking head with independent head and eye movements was used as a talking agent interacting with a user during a simple card game offering different levels of help and guidance. We analyzed the user performance and how the quality of assistance given by the embodied conversational agent was perceived. The experiment showed that users can profit from its presence and its facial deictic cues.


P1-3 Text Input Disambiguation Supported on a Hierarchical User Model
Bento, Carlos - Gil, Nuno
As formated for the printed proceedings - 24.ps - 24.pdf - pages 253-258
As delivered by the authors - 24_pdf_file.pdf
Abstract :
Mobile phones are used for various tasks that go far from voice communication. A popular use is for composition of short messages (SMSs), but other applications are also available like email, agenda, contact, and note management. All these uses have in common the need for text input on a small keyboard with ambiguity problems.
Various techniques are currently used for input disambiguation, with variable results in terms of usability and efficiency. Some techniques achieve good performance with messages composed by words from a dictionary but poorly when a significant number of words are not in memory.
In this paper we present a solution for text disambiguation supported on a general model for disambiguation plus a user model generated from previous messages sent by the user. We present results obtained with this approach and discuss future improvements.


P1-4 Supervised Learning of an Abstract Context Model for an Intelligent Environment
Brdiczka, Oliver - Reignier, Patrick - Crowley, James L.
As formated for the printed proceedings - 8.ps - 8.pdf - pages 259-264
As delivered by the authors - 8_pdf_file.pdf
Abstract :
This paper addresses the problem of supervised learning in intelligent environments. An intelligent environment perceives user activity and offers a number of services according to the perceived information about the user. An abstract context model in the form of a situation network is used to represent the intelligent environment, its occupants and their activities. The context model consists of situations, roles played by entities and relations between these entities. The objective is to adapt the system services, which are associated to the situations of the model, to the changing needs of the user. For this, a supervisor gives feedback by correcting system services that are found to be inappropriate to user needs. The situation network can be developed by exchanging the system service-situation association, by splitting the situation, or by learning new roles. The situation split is interpreted as a replacement of the former situation by sub-situations whose number and characteristics are determined using conceptual or decision tree algorithms. Different algorithms have been tested on a context model within the SmartOffice environment of the PRIMA research group. The decision tree algorithm (ID3) has been found to give the best results.


P1-5 An ontology-based context management and reasoning process for UbiComp applications
Christopoulou, Eleni - Goumopoulos, Christos - Kameas, Achilles
As formated for the printed proceedings - 14.ps - 14.pdf - pages 265-270
As delivered by the authors - 14_pdf_file.pdf
Abstract :
UbiComp applications operate within an extremely dynamic and heterogeneous environment and have to dynamically adapt to changes in their environment as a result of usersí or other actorsí activities. So context definition, representation, management and use become important factors that affect their operation. To ease the development of such applications it is necessary to decouple application composition from context acquisition and representation, and at the same time provide universal models and mechanisms to manage context. In this paper is presented an approach for building a context-aware UbiComp system organised in hierarchical levels. The focus of the paper is on an ontology-based context modelling, management and reasoning process developed for composing context-aware UbiComp applications from AmI artefacts.


P1-6 Multimodal Appliance Cooperation based on Explicit Goals : Concepts and Potentials
Heider, Thomas - Kirste, Thomas
As formated for the printed proceedings - 76.ps - 76.pdf - pages 271-276
As delivered by the authors - 76_pdf_file.pdf
Abstract :
Smart Environments are increasingly composed from individual components ("smart appliances") that have to assemble themselves into a coherently acting ensemble. This requires software technologies that enable appliances to cooperate spontaneously on behalf of the users needs.
In this paper we will illustrate why a goal based approach is necessary and how explicit goals can be used to find system comprehensive strategies and how goals can be used as a benchmark to evaluate the system design.


P1-7 Peis Ecologies : Ambient Intelligence meets Autonomous Robotics
Saffiotti, Alessandro - Broxvall, Mathias
As formated for the printed proceedings - 29.ps - 29.pdf - pages 277-282
As delivered by the authors - 29_pdf_file.pdf
Abstract :
A common vision in the field of autonomous robotics is to create a skilled robot companion that is able to live in our homes and perform physical tasks to help us in our everyday life. Another vision, coming from the field of ambient intelligence, is to create a network of intelligent home devices that provide us with information, communication, and entertainment. We propose to combine these two visions into the new concept of an ecology of networked Physically Embedded Intelligent Systems (PEIS). In this paper, we define this concept, and illustrate it by describing an experimental system that involves real robotic devices.