![]() |
ESCA - NATO/RSG 10 Workshop on
|
![]() |
In the past we have believed that, in an embedded system framework, the best way of solving the problem of an on-board interaction system was to make a judicious choice of the communication channel and to improve its related techniques. That was the very case for the speech processing devices that are to outfit the next generation of manned combat aircraft. In France, a similar choice has been made for the Rafale D. What is getting strange right now, is that we realize that the only use of speech processing systems is not sufficient although an on-board aircraft operationnal use is more and more feasable. We also realize, considering the pilot's associate for instance, that man-machine interface and dialogue are getting more and more complex and cannot be based on one single channel. In other words, speech processing although interesting by itself appears to be used with more benefit if we are able to "multiplex" it with other communication devices. In the context of a pilot's assistance, it is becoming necessary to envisage a multimodal approach. But in this way, we have to tackle new questions : what is an interaction ? how to get to represent the semantics of an interaction? etc... Finally we have to define a new global conceptual approach which could be used to model a multimodal man-machine interaction integrating voice. In this paper we describe some essential elements about this approach and introduce some adequate notions.
Bibliographic reference. Bares, Michel (1993): "Conceptual approach to model the man-machine interface: (application to the pilot's associate)", In AST-1993, 71-74.