부산대학교 도서관

The role of recognition and imitation of human actions in humanoid robots is very important, but there is a problem of inaccurate recognition and imitation. Ordinary methods cannot solve the problem of recognizing and imitating human movements in humanoid robots, and the methods are unreasonable. Therefore, this paper proposes a machine learning to recognize and imitate human motion analysis. First of all, artificial intelligence is used to inspect the robot, and the indicators are divided according to the requirements of recognizing and imitating human actions to reduce them Identify and imitate interference factors in human movements. Then, artificial intelligence detects the humanoid robot to recognize and imitate human actions, forms a recognition and imitation of human action schemes, and recognizes and imitates human action results Conduct a comprehensive analysis. MATLAB simulation shows that under certain evaluation criteria, machine learning can recognize and imitate the accuracy of human movements of humanoid robots The time to recognize and imitate human movements is better than ordinary methods.