Gradient Based Identification of Wiener Systems

İngilizce (OpenAlex)

Insystems consisting of a single block, the establishment of feedback control andinput-output relationship is common in classical control. However, consideringnatural systems, there are many block structures and non-linear functionswithin these blocks. The subject of this study is the identification of Wienertype systems, which is the system structure in which the input signal isapplied to the linear block and its output is taken from the nonlinearfunction. It is possible to use different control algorithms withidentification in this type of system where state feedback control is notpossible. In system identification, the mathematical relationship between theinput and output of the system has been obtained by the choice of autoregressive with exogenous inputs (ARX)-polynomial cascade connection with usingthe least squares method and gradient information. Three different benchmarksystems have been studied in MATLAB/Simulink and the performances of theidentifications made with the dataset are presented graphically.