Investigation On Effect Of Error Order Selection In System Identification

Mohd Bakri, Mohamad Syahmi (2017) Investigation On Effect Of Error Order Selection In System Identification. Project Report. UTeM, Melaka, Malaysia. (Submitted)

Text (24 Pages) Investigation On Effect Of Error Order Selection In System Identification.pdf - Submitted Version Download (502kB)

Abstract

System identification is a technique or application that aims to develop mathematical models for dynamical systems using measured input and output signals. There are 4 important steps which is the observed data, model, parameter estimation and validation. In this project, model structure selection is important because the objective is to find the effect of the error order selection on system identification. Linear model which is ARX and ARMAX model is used to make the simulation to investigate effect of error order. This project was carried out using the Graphical User Interface (GUI) in MATLAB application. Firstly, 3 equations were randomly created with supervisor guide and data acquisition was run to generate 500 data by using the MATLAB software command and save it into folder for next step use. After that, system identification toolbox is open and imports the data that generate from the equation. Next step is modelling where the data use ARMAX model and different sequence in error order or . This project result is based on 5 performance indicator like Means Square Error (MSE), Model Output, Model Residuals, Akaike’s Final Prediction Error (FPE) and Parameter Values. This entire indicator will pop out after click on button indicator or from the data of each model. Based on result that acquire from this investigation, the effect or error order selection have a very small difference compared to the true model. This is maybe because the data was not completely scatter and need to bigger. So the conclusion is, there is small effect of error order selection in system identification.

Actions (login required)

View Item