In the past years spoken dialogue systems became more and more sophisticated, i.e., they allow for rather complex and flexible dialogues. On the other hand the functionality of devices, applications and services (henceforth applications) and the amount of digital content increased rapidly. Due to the network age (Internet, personal area networks,...) the applicationswecan/want to control change dynamically as well as the content. Therefore a dialogue system cannot be anymore manually designed for one application before product release. Instead a layer in between is required, which translates the device functionalities in a way that they can be used via a more or less generic dialogue system. Furthermore, a network of devices creates new functionalities not possible with a single device. Users will have complex wishes; often several devices will be necessary to solve such a wish. However, the user in most cases would not like to care about how many and exactly which applications are used to fulfill his wish. From his perspective he sees the functionality provided by a network of devices as a virtual device. In this paper a module is described which builds a link between dialogue system and an ensemble of applications. Functionalities of applications are described in a formal way. A planning component searches for the applications necessary to solve a complex user wish and the sequence of action which has to be performed. Simplicity of the approach is very important. the module has been successfully integrated in a prototype system which was publicly demonstrated.
|Number of pages||4|
|Publication status||Published - 2002|
|Title||7th International Conference on Spoken Language Processing|
|Abbreviated title||ICSLP 2002|
|Duration||16 - 20 September 2002|