System Identification for SISO Systems

System identification is a well-established field, grew both in size and diversity over the last several decades. In addition, system identification methods can handle an extensive range of system dynamics without knowledge of the actual system physics. In this report, system identification for single-input and single-output (SISO) system and the improvisation techniques are discussed. The most significant criteria in system identification are selection of suitable model structure, excitation signal, signal to noise ratio (SNR) and frequency. This can be done by using System Identification Toolbox in MATLAB, where it will build an accurate and simplified model from complex system with noisy time-series data. Three different systems are discussed by using ARX, ARMAX and OE models. For each system, five case studies with different orders are discussed. In addition, different types of excitation signals are used in order to get the best results. The model fitting, bode plot, step response and residual plot are obtained by using System Identification Toolbox. Besides that, the mathematical equations which are used to calculate the parameters are also presented in the following section. Based on the fitting, the best models are interpreted. The results of each case study show the importance of model selection for different scenarios.