Bianca Angelini, Giuliano Antoniol, Massimo Dal Zotto, Renato De Mori, Diego Giuliani, Roberto Gretter and Gianni Lazzari
This paper describes the first version of a system that aims to become both a demo prototype of a robot-operator dialogue component and a starting version of a probabilistic language model acquisition environment. A simulation has been carried out, in which a mobile robot, guided or teleguided by an operator, is able to orientate itself all along the corridors of our institute and to carry out simple navigation tasks. The most important functions of the dialogue, in this simulation, are the definition of goals and the possibility of avoiding obstacles or passing through critical points.
The spoken dialogue system is composed of three modules: an automatic speech recognizer, a text-to-speech synthesizer and a component implementing the spoken dialogue model.
Two components constituting this system will be described and discussed, that is the recognizer and the dialogue model. Preliminary recognition performance of the system is also presented.