Non linear system identification using kernel based exponentially extended random vector functional link network

Abstract

Identification of nonlinear systems finds extensive applications in control design and stability analysis. To identify complex nonlinear systems, the neural network has drawn the attention of many researchers due to its broad application area. In this paper, an improved identification method based on Kernel Exponentially Extended Random Vector Functional Link Network (KERVFLN) has been proposed for nonlinear system identification. Good generalization capability, fast learning speed, simple architecture and the direct connection between input and output nodes along with non linear enhancement nodes with random weights of traditional Random Vector Functional Link Network (RVFLN) are very essential to industrial applications. To avoid the selection of the number of hidden nodes and hidden mapping function, kernel function has been used in this paper to increase the stability. The input is extended using trigonometric expansion which increases the accuracy of the algorithm when ever there is a sudden random change. In case of KERVFLN the number of enhancement nodes and its corresponding activation function need not to be known if its corresponding kernel function is given. To verify the accuracy of the proposed model, some benchmark Monte Carlo simulations and one SISO system are carried out through simulation study and the obtained results are compared with some established techniques such as original RVFLN, Extreme Learning Machine (ELM), and Least Mean Square (LMS). The efficiency of the proposed technique has been tested with the real time data set as well. The prediction accuracy of the proposed method KERVFLN is higher than the normal RVFLN for different nonlinear systems which is clear from the performance evaluation section.