ABSTRACT
The introduction of speech recognition in an air traffic control (ATC) simulator is discussed. Supporting the speech recogniser with situational knowledge is identified as a strategy to improve recognition performance. A model of user behaviour is proposed as a means to generate an estimation of which actions the controllers are likely to take next. During that estimation, both situation-specific and user-specific influences have to be considered. Experiments are discussed that help to obtain the relevant data about expert behaviour.
