Veronique Daurensan

Adaptive user interface for process control based on multi-agent approach

Teams of operators are required to monitor and control complex real-time processes. Process information comes from different sources and is often displayed by existing User Interfaces using a variety of visual and auditory forms and compressed into narrow time-windows. Most presentation modalities are fixed during interface design and are not capable of adaptation during system operation. The operators alone must provide the flexibility required in order to deal with difficult and unplanned situations. This paper presents an innovative Auto-Adaptive Multimedia Interface (AAMI) architecture, based on Intelligent Agent collaboration, designed to overcome the above drawbacks. The use of this technology should speed up the design and the implementation of human-centred multimedia interfaces, and significantly enhance their usability. The proposed architecture separates generic knowledge about adaptive user interface management from application specific knowledge in order to provide a generic framework suitable to be customised to different application domains. Benefits from the AAMI approach are evaluated by developing two industrial field-test application including Electrical Network Management and Thermal Plant Supervision system. The paper reports the architecture and the basic design principles of the generic framework as well details of the two applications. The work is being carried out within the European ESPRIT project: AMEBICA.