PARAMETER ESTIMATION OF MATHEMATICAL MODELLING USING HYBRID EXTENDED KALMAN FILTER

• Main Author: Putri Gurusinga, Sadrika
• Format: Final Project
• Language: Indonesia
• Subjects: Matematika
• Online Access: https://digilib.itb.ac.id/gdl/view/33786
• Institution: Institut Teknologi Bandung

A mathematical model may contain one or more parameters. Their values can be determined from literature or estimated based on experimental data. In the latter case, we face parameter estimation problem. There are numerous kinds of method that can be used for parameter estimation. Here we use Hybrid Extended Kalman Filter. In the first part, we focus on how Kalman Filter can be used to estimate a state value based on observed data when the mathematical model is of discrete type. In the second part, we deal with continuous model, hence Hybrid Kalman Filter is used. We show that both methods are able to estimate the state of equations, not only in simple model but also in complicated model. In order to be used for parameter estimation, this research used Hybrid Extended Kalman Filter since the mathematical model being considered becomes nonlinear model. We show that using Hybrid Extended Kalman Filter, parameter in a mathematical model can be well estimated. Hence, it can be an alternative method for parameter estimation in mathematical modelling.