Controlling Collaborative Robots in Factory Environments with Turkish Voice Commands

This study aims to control multiple collaborative robots with Turkish voice commands from a single source. For this purpose, a software architecture has been proposed within the scope of the study. The proposed architecture includes processes such as noise reduction, speech-to-text conversion, command generation, and execution of commands on the robot interface. Additionally, a prototype application of the proposed architecture has been developed. According to expert opinion, DeepFilterNet3 has been used for noise reduction operations in the developed prototype application, yielding successful results. The OpenAI Whisper model has been improved and utilized for the speech-to-text conversion process. The text obtained from the OpenAI Whisper model has been transformed into commands to be executed on the robot interface, and it has been successfully operated by the collaborative Universal Robot 5. As a result, it has been observed that the resulting sequential program functions effectively.