Real-time Humanoid Robot Dance System Based on Music Genre Classification
Swan Lu, Han‐Pang Huang, Ze-Feng Zhan · Research Square · 2023
Abstract Nowadays, people's pursuit of artificial intelligence is not limited to the scope of convenience or intelligence. The entertainment use of robots is gradually being considered and used, and the dance of humanoid robots is one of people's pursuit of robots imitating humans and the entertainment of robots. This research uses a real-time beat tracker and a unique humanoid robot dance step designer to design a real-time humanoid robot dance system. Via Ratio was proposed for robot dance, with a real-time beat tracker and music speed estimator. For music, neural networks were used for music genre classification and the GTZAN music dataset was used for the training. In the design of movements, corresponding dance movements will also be designed according to different styles. Some dance motions were designed and make our humanoid robot NINO, which was developed by our lab, success to dance with the beat of the music. Also, the robot can identify the genre and choose the motion corresponding to the music to dance. The experiments for the proposed system and functions were carried out on a life-sized humanoid robot NINO. The results showed that this system can successfully make the robot dance to music and correctly identify the music genre. The main contribution of this research is a dance system designed for a life-sized humanoid robot. Also, some music elements such as genre and rhythm were combined into the dance system. The robot can dance to the beat with Via Ratio and beat tracker. The music genre classification system was designed to achieve the purpose of robot dance being gorgeous, artistic, and not rigid.