A VMD Based Extreme Learning Machine Approach for Nonlinear System Identification
Shaktinarayana Mishra, Prachitara Satapathy, Lokanath Tripathy, P.K. Dash · 2019
Now-a-days, most of the techniques applied to identify the non-linear system are based on machine intelligence. In this work, a new Varitional Mode Decomposition based Extreme Learning Machine (VMD-ELM) is proposed to identify various non-linear systems. The VMD technique helps to denoise the input and output of the unknown non-linear systems before fed to the ELM network. The proposed VMD-ELM method helps to achieve more precise identification result with smaller computation complexity. The superior performance of the proposed VMD-ELM method is verified by comparing it with basic ELM and Emprical Mode Decomposition based ELM (EMD-ELM) for various non-linear systems in MATLAB/script environment. The efficacy of the proposed method is further validated through dSPACE DS1104 for a practical boiler system.