Abstract

This paper describes an application of Automated Speech Processing in two food ordering environments: drive--in restaurants and ordering by telephone. After describing a general model for such dialogues, the two environments are described in detail: covering word use and dialogue structure. Two prototypes have been developed for these environments. Testing these prototypes revealed that additional training of the speech recogniser is essential is this domain. Finally, a simulation study was done in which the human operator was replaced by a simulated system making use of a minimal set of responses selected from the recorded corpus. The simulation study showed that more or less standard dialogues can be simulated with such a system.