This article describes the Philips automatic train timetable informati
on system which enables the user to call up accurate information about
train connections between 1200 German cities over the telephone. In c
ontrast to most of the inquiry systems available so far, the caller ca
n talk to our system in unrestricted, natural and fluent speech, very
much like talking to a human operator. No instructions are given befor
ehand. The system consists of four main components: speech recognition
, speech understanding, dialogue control, and speech output. They are
separated into independent modules and executed sequentially. The spee
ch recogniser creates a word graph from the spoken input. This word gr
aph is then passed to the understanding component which computes the m
eaning, using an attributed stochastic context-free grammar. A dialogu
e manager analyses the results and either accesses the database or com
es up with another question if necessary. The system has been made ava
ilable to the general public in an ongoing field test, both to gather
speech data and to evaluate its performance.