SYSTEM IDENTIFICATION FRAME WORK USING NEURAL NETWORKS AND EXTREME LEARNING MACHINE

Abstract

System identification is introduced with machine learning as to handle the problem of simulation and control design, also the algorithm of identification is applied in a small embedded board to verify its mission of having a device which collect the data then start the identification process, after that the controller design is the part of the new needed task. Different learning techniques using neural networks are showed to obtain a suitable algorithm for the identification process. The suitable algorithm for this work mean low processing power, high accuracy and low processing time. The results from the neural network were showed and verified in comparison with the real data collected from different dynamical systems.