A mechanical vibration identification method for sub-Nyquist sampling based on ESPRIT-like algorithm

This paper proposes a mechanical vibration identification method for sub-Nyquist sampling in theory, through the study on the ESPRIT algorithm. Then the method is improved for the problems encountered in the practical application: the big computational amount of solving generalized eigenvalues, the aggravation of ill-condition brought by correlation matrices, the big errors of attenuation coefficients. The improved identification method can be used to practically identify the sub-Nyquist sampling of mechanical vibrations. And the vibration signal identified can be recovered accurately. The validity of the method is verified through the simulation.