Affiliations: [a] Department of Electronics and Multimedia Communications, FEI Technical University of Košice, Košice, Slovakia | [b] Institute of Informatics Slovak Academy of Sciences, Bratislava, Slovakia | [x] Laboratory of Speech Acoustics, Department of Telecommunications and Media-Informatics, Budapest University of Technology and Economics, Budapest, Hungary | [y] Dipartimento di Psicologia and IIASS, Seconda Università di Napoli, Vietri Sul Mare, Salerno, Italy
Correspondence:
[*]
Corresponding author: Stanislav Ondas, Department of Electronics and Multimedia Communications, FEI Technical University of Košice, Košice, Slovakia. E-mail: stanislav.ondas@tuke.sk
Abstract: The proposed paper deals with the design and development of SIMONA – the Slovak embodied conversational agent. It was created as an extension of a multimodal dialogue system with the BML realizer Elckerlyc as an embodied agent. Elckerlyc is a BML compliant behavior realizer developed by the Human Media Interaction group at the University of Twente. SIMONA is an asymmetric multimodal system, with which the user communicates by spoken utterances and the system responds using speech, gestures, head movements, and body posture. Galaxy hub-server architecture was adopted in order to construct an agent based system in which agents are configured in a star topology and they cooperate upon the common data space in a form of database with several pools. Servers communicate together in two-level communication architecture. A speech recognition system for the Slovak language was included. For producing speech output, text-to-speech engine was prepared and connected to the agent through the MS SAPI interface. The conversion between Slovak visemes and generalized SAPI 5 visemes was defined. A new distributed dialogue manager was designed to enable event-driven, task-oriented interactions that can be controlled in several modes, e.g. finite-state and frame-based ones. The initiative in the interaction can be on the side of the system, of the user, or they can alternate in leading the dialog.
