Automatic Speech Recognition for Human-Robot Interaction on the Humanoid Robot Barelang 7

Abstract

This paper proposes a Human-Robot Interaction (HRI) system through voice commands. The study used the Barelang 7 humanoid dancing robot platform comprising 29 Degrees of Freedom (DoF). Automatic Speech Recognition (ASR) was developed using Python programming language and the VOSK Offline Speech Recognition toolkit to recognize five instructions spoken by random speakers. The voice recognition results in ASR are then sent to the robot via serial communication to realize movements according to spoken instructions. From tests conducted on five speakers randomly with 100 attempts, a Word Error Rate (WER) value of 0.096 was obtained. Meanwhile, testing the entire HRI system achieved a success rate of 92%. These results indicate that the HRI system has been successfully implemented on humanoid robots, although there are still some errors in the ASR system, that affect the overall performance of the HRI system. The results of this research contribute to the research and development of HRI systems, especially in humanoid robots, which are still not widely studied.