On Semantic Graph Language Processing for Mobile Robot Voice Interaction

Abstract

This paper describes a simple semantic graph based model for processing natural language commands issued to a mobile robot. The proposed model is intended for translating natural language commands given by naïve users into an action or sequence of actions that the robot can execute via its available functionality, in order to complete the commands. This approach to language processing is easily extensible through automated learning, it also is simpler and more scalable than hard-coded command to action mapping, while also being flexible and covering any number of command formulations that could be generated by a user.