Gestaltung avatarbasierter natürlichsprachlicher Hillfesystme für den Einsatz in Fahrzeugen (Volume 20) (English shop)

Preview

Table of Contents, Datei (71 KB)
Extract, Datei (110 KB)

Drivers of modern vehicles require access to information about rarely used or hidden functions. An innovative approach to providing this information even while driving is a natural-language, avatar-based in-vehicle help system.

First, a reference architecture for avatar systems is developed and applied to the automotive domain. The software implementation consists of various modules providing functions for speech input and output, reaction determination, real-time 3D rendering and responses to messages from the bus systems.

For the development of dialogues, an authoring tool is developed that enables graphical modelling. This tool is part of a seven-phase procedure model for transferring structured information into dialogues. Finally, the avatar system and the multimedia content are examined in the vehicle with 67 test subjects.

Dr. Valentin Nicolescu studied economics and business administration at the Universität Hohenheim, completing his studies with a diploma thesis on avatars on the Internet. From 2003 to 2008 he worked as a research associate at the Chair for Information Systems at the Technische Universität München. There, as a technical consultant at the SAP University Competence Center, he worked on areas including ABAP development, SAP Process Integration and SAP NetWeaver Portal. From 2006 he developed, in a research project within the INI.TUM framework programme together with Audi, the prototype of an avatar system in a test vehicle, which formed the basis of his dissertation. Since 2009 he has been working as an IT Development Consultant at Webasto AG in Stockdorf.