Inclusión de la lengua Wayuunaiki en el reconocimiento de comandos de voz del robot social Pepper empleando la metodología de transformación de modelos

Abstract

The present proposal seeks to provide the Pepper social robot, which the Santo Tomás University has, with the recognition capacity for voice commands in the Wayuunaiki language. In this way, the use of more advanced social robotics functions such as assistance and service will be allowed. Likewise, through the expected results and products, the aim is to reduce the technological gap in the Wayuu community by providing new capabilities to the Kailumá project of the Santo Tomás University. The voice recognition tools available are based on technologies that require a large amount of previous data along with a training process for these tools. This means that the language support by these tools is given for languages ​​such as Spanish and English, but not for languages ​​with few speakers, as is the case of Colombian indigenous languages. Therefore, a solution to this problem is proposed by using model transformation chains. For this, voice commands in the Wayuunaiki language are modeled, which will be the input of the transformation chain; thus, it is required to configure, train and model a speech recognition tool that will serve as the output of the transformation chain. The generation of such models arises from the analysis of the predefined set of voice commands for social robotics for a domestic service environment. The transformation chain will be implemented as a ROS node to be enabled in the Pepper robot