Generating natural-language like robot motion command set through machine learning

Abstract

A learning user interface (LUI) for commanding robot motions in a natural-language like form is presented. Unlike other natural language interfaces, it has the ability to learn how to execute a motion command from the demonstration given by the user. An effective learning algorithm is developed by restricting the learning domain to motions and spatial relationships. The LUI has the ability to generalize the concept learned from a demonstration to other situations. The LUI allows the robot operator to issue qualitative, natural-language like motion commands rather than precise numeric commands involving coordinate specifications. This capability speeds up robotic operations in many applications, especially in unstructured environments such as hazardous waste clean-up and underwater exploration.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>