A Survey of Resonance Frequency Detection in Robot Joint Systems

Abstract

The transmission mechanism of the robot joint system is usually composed of a reducer, a coupling, etc., which inevitably introduces elasticity into the system, causing mechanical oscillations. In order to effectively suppress the mechanical oscillation phenomenon, the detection of the oscillation frequency is essential. This article models the robot joint and analyzes the resonance mechanism, analyzes the basic idea of each resonance detection method of the joint system. To solve the common problem in joint servo system, methods including sliding DFT, windowed bispectrum interpolation FFT, wavelet transform method, Welch method, Kalman filter detection method and neural network-based resonance detection have been proposed successively. This article compares the advantages and disadvantages of related methods, summarizes the existing research results and problems to be solved, and looks forward to the future development trend and application prospects of detection technology.