Multilingual verbal interaction between humans and robots - Modeling and Implementation

Abstract

Language barriers are one of the biggest challenges that humans in general and children in particular face in their verbal interactions. Many reseach projets with mutli-countries scope are expected to give a significant help to teachers and doctors to support children with verbal communication disabilities by using humanoid robots. The usage of robot agent is excpected to stimulate the verbal communication based on natural language interactions with this kind of children. Such stimulation may be of great importance in improving the ability of these children to interact verbally with their teachers and classmates. In this context, we propose through this work to extend the language capabilities of humanoid robots to be able to communicate with children using differnet natural languages. This should allow a humanoid robot to recognize the spoken or written message of each children group and thus use this language to interact with. In this paper, we present the design of a generic models and the implementation of a software components to allow language neutral verbal interactions between children and humanoid robots. An extended ontology models will be provided to serve as a knowledge database of a set of concepts and theirs propoerties to represent the semantic aspects describing the relations between human-robot vebal communications. Based on thus generic models, we will provide an optimized new algorithms and their Python implementation to extend the natural language processing Python pakages. In this paper, a set of human-Robot verbal ineractions models are presented and the resultant conclusions and perspectives are discussed