FLUIDS Project Partners:
MIZAR Automazione S.p.A. (Co-ordinator)
Universidad Politécnica de Madrid
DFKI German Research Center for Artificial Intelligence GmbH
Consorzio 5T
The need for interaction between an automated control system and its users often exceeds the capabilities of current interfaces and calls for the study of advanced software environments to support the design and development of interfaces showing intelligent behaviour. The need to increase and improve user-system interaction is particularly strong in real-time decision support systems.
In order to obtain a satisfactory degree of involvement between the user and the decisions he/she takes an interface is required for which two aspects have to be carefully designed:
The main objective of the FLUIDS project is the design of a general environment for building intelligent interfaces of real-time decision support systems, where a model of the operator-system conversation is supported by providing varied answers and explanations that are elaborated using artificial intelligence techniques and presented to the user by integrating advanced multimedia techniques.
The FLUIDS project aims to provide two main results:
The project is being developed in two steps:
The two main features of the new generation of user interfaces for real-time decision support systems developed in the FLUIDS project are:
The question answering capability of the FLUIDS approach resides in a general framework of structured knowledge for real-time decision support that can be instantiated for any concrete application domain. This framework identifies the different knowledge elements occurring in a generic decision making scenario and links them with their corresponding roles in each step of the decision support process. Moreover, the generic model includes knowledge to associate potential queries of a human decision maker with appropriate problem solving methods which are needed to derive suitable answers to these questions. The KSM (Knowledge Structure Manager) environment was selected by FLUIDS as the tool for supporting the approach based on knowledge units for the formulation of structured knowledge models.
The capability to flexibly tailor presentation to users's needs is one of the key aspects of the intelligent multimedia component. Handling dynamic non deterministic streams of application data (the typical working condition of a real-time system) requires also dynamic adaptation of interaction styles; multimedia dialogues are considered as a sequence of communicative acts to achieve certain goals. With the solution adopted in FLUIDS, multimedia interaction is regarded as a non-routine design task and a uniform planning approach is used for the knowledge-based design of multimedia dialogues. The current version of the FLUIDS development environment supports multimodal input by combining (1) menu-based natural language for the formulation of questions, and (2) deictic pointing gestures to select potential problem areas as optional for ``what happens?'' queries. On the output side, various media combinations are employed including natural language text, speech, graphical representation and animation.
FLUIDS offers an innovative architecture and a structured approach for constructing intelligent interfaces for real-time decision support. Its main features are the advanced KSM methodology for knowledge engineering and sophisticated plan-based approach to multimedia interaction rooted in well established linguistic theory. The technical realisation strategy for the FLUIDS software environment exploits new emerging paradigms in computing to maximise the potential for broad and successful applications.
The first year of the project, which coincides with phase I of the project work plan, has been concluded positively by providing a working first version of the FLUIDS development environment. This was demonstrated to the Commission in October 96. An off-line demonstrator was also built and is now being used for the first cycle of user consultation. Revisions of the software environment have taken place and plans for the next implementations of the FLUIDS interfaces are in progress.
A report on the market potential for the Intelligent User interface has been compiled giving the market dimension and prospectives of the FLUIDS product in different application domains.
Encouraging results have also been achieved by the project with the decision to focus first year efforts also on the dissemination activity with the objective to promote the practical implementations of the Artificial Intelligence techniques to ``real world'' systems and thus challenge the scepticism that is often present in the scientific community with regard to Artificial Intelligence. The consortium have set up the FLUIDS Web Site (http://www.dfki.uni-sb.de/fluids/) where all the documents produced by FLUIDS are available, together with other additional information.
The mission of the FLUIDS project is to provide a general software
environment for building intelligent interfaces for decision support
systems and a methodology for designing interface models. Although the
approach is being tested for the specific domain of road traffic
management, the goal is not just to solve this particular application
problem but to offer a generic solution suitable for any real-time
control system. Potential application scenarios for intelligent
multimedia interfaces for real- time decision support systems are the
fields of: traffic management for non-road transport,
telecommunication, energy, emergency management, health care, control
in an industrial environment etc.