NTNU Cyborg With Communicational Abilities

Abstract

Motivated by the unlimited possibilities of combining autonomous robots with speech recognition and synthesis, this thesis aims to provide a robot with the abilities of communicating with human beings in its surroundings. There exists a variety of such solutions within today s technology, utilizing different kinds of technology making it possible for robots to interact with humans through speech. Suitable tools and frameworks are examined aiming to implement a solution providing the robot NTNU Cyborg with conversational skills. Such a solution must solve the challenges of converting speech into text, generate appropriate answers to associated questions and finalize the communication through speech synthesis. Utilizing the functionality provided by the Robot Operating System in conjunction with the final solution of this thesis, made the robot capable of communicating with human beings in its surroundings through a microphone and an Ipad. Several platforms for the purpose of speech recognition were assessed, whereas Google was the platform assigned the highest priority. To achieve a reasonable conversation taking place between a human being and the robot, software for generating answers to corresponding questions was crucial. Platforms acting as chatbots were combined with functionality implemented for the purpose of making a conversation perceived as interesting and humorous. The resulting reply generated by the robot was combined with existing software for converting text into speech, forming the robot s face. That is, the software Brain Language. A satisfactory response time and accuracy of the conversations taking place with the robot was obtained. Such conversations include humorous and informative answers in addition to functionality forming a personality for the robot when interacting with it through speech.