Modeling and development of a spoken natural language interface for autonomous robot interaction

Human-robot interaction with autonomous robots, based on spoken language, is an user intuitive approach to unlock further developments in the field of robotics, although still a challenging problem. The focus of the paper is at developing a spoken natural language dialogue interface, flexible and scalable in terms of number of understood and executed commands, without environment noise constraints. Our main objetive is to perform a word morphological analysis in order to provide a text classification. The developed system uses Google Speech Recognizer installed in a smartphone and a laptop connected to it via Bluetooth interface. The proposed algorithm is capable to recognize words despite grammatical tense, with a low error rate and a high language flexibility. The technologies and devices applied guarantee a light, efficient and low-cost solution, unlocking new possibilities for further study and development of better, more autonomous and environment-adaptable robots.