The aim of this thesis is the development of a Spoken Dialogue Manager (SDM) for Intelligent Environments. The SDM is able to manage, smoothly and naturally, mixed-initiative dialogues about multiple interdependent tasks. It resolves task ambiguities and conflicts and allows users to interrupt and resume tasks. The multitasking nature of spoken dialogues within Intelligent Environments is one of the main challenges since the switching among tasks may depend on the user input, the input from external entities, or on the task definition itself. In case of conflicts related to overlapping dialogue domains, a problem solving dialogue can be initiated. Moreover, based on the various tasks and the dialogue domains involved, the SDM is able to guide the user’s navigation through the inter- or independent spoken dialogue domains.

Several evaluation sessions have been carried out to proof the concept, to find out how users cope with a multitasking spoken dialogue, and to investigate how the model-based approach fosters the recognition and more specifically the understanding rate of the system. The thesis is carried out within the framework of the EU-funded Project ATRACO (Adaptive and TRusted Ambient eCOlogies ATRACO).