Intelligent Fault Diagnosis Method of Robot Joint—Harmonic Reducer Based on EMD and SVM

The harmonic reducer is a vital component of a robot, responsible for high-precision and high-torque transmission. It finds wide applications in the field of robotics. However, due to its complex internal structure and the generation of significant friction and heat during operation, it is prone to failures that can significantly impact the robot’s performance. To address this issue, we propose a fault diagnosis method based on empirical mode decomposition (EMD) and support vector machine (SVM). The proposed method involves several steps. First, the fault data are decomposed using EMD, and the resulting intrinsic mode functions (IMFs) are evaluated based on their steepness values and correlation with the original signal. Only the most relevant IMFs are selected for reconstruction. Next, the reconstructed signal is band-pass filtered according to the recommended centre frequency and bandwidth parameters. The filtered signal is then demodulated using square envelope analysis. Then, time-domain and frequency-domain features of the demodulated square envelope signal are extracted and used as inputs for training and testing an SVM classifier. The SVM classifier learns to associate these features with specific fault types or normal operations, enabling fault identification. Finally, a comparative experiment is conducted to evaluate the performance of the proposed method. The results demonstrate that the method achieves high accuracy in fault identification for the harmonic reducer. This approach provides theoretical support for ensuring the safe operation of both the harmonic reducer and the robot as a whole.