University of Arkansas, Fayetteville


In the very near future, as pervasive computing takes root, there will be an explosion of everyday objects that are uniquely identifiable and wrapped by a computational layer- effectively bringing the object to life. An important component of this system is the mechanism that will allow humans to interface with the objects. Menu Based Natural Language Interfaces (MBNLI) seem like a good candidate for this job because of the intuitive way in which they allow the user to build commands. However, the MBNLI system will have to scale with the number of objects in the system. This project describes context free grammar modules which are small grammar files that can be composed to form a larger grammar. Grammar modules can then be associated with individual objects, and in this way allow the MBNU to scale according to the size of the system.