UMP Institutional Repository

How Important the Error Covariance in Simulated Kalman Filter?

Nor Hidayati, Abd Aziz and Zuwairie, Ibrahim and Saifudin, Razali and Bakare, Taofiq Adeola and Nor Azlina, Ab. Aziz (2016) How Important the Error Covariance in Simulated Kalman Filter? In: Proceedings of The National Conference for Postgraduate Research (NCON-PGR 2016), 24-25 September 2016 , Universiti Malaysia Pahang (UMP), Pekan, Pahang. pp. 315-320..

[img]
Preview
PDF
P044 pg315-320.pdf

Download (678kB) | Preview

Abstract

The process of searching good parameter values is a non-trivial task for metaheuristic algorithms. When two algorithms are comparable in terms of speed and probability of convergence, the algorithm with less number of parameters is always preferred. This paper discussed the importance of the initial error covariance parameter, P(0), in Simulated Kalman Filter (SKF) with an intent to make SKF a parameter-less algorithm. To evaluate the importance of initial error covariance value in SKF, several values were selected and statistical analyses using nonparametric Friedman and Wilcoxon signed rank tests were carried out to see if different initial error covariance has any significant difference in the final outcome. The results prove that no matter what the initial error covariance is, SKF algorithm still managed to converge to near-optimal value without any significant degradation or improvement.

Item Type: Conference or Workshop Item (Lecture)
Uncontrolled Keywords: SKF; Kalman; error covariance; metaheuristics; optimization
Subjects: T Technology > TK Electrical engineering. Electronics Nuclear engineering
Faculty/Division: Faculty of Electrical & Electronic Engineering
Depositing User: Noorul Farina Arifin
Date Deposited: 30 Sep 2016 08:13
Last Modified: 08 Feb 2018 02:49
URI: http://umpir.ump.edu.my/id/eprint/14582
Download Statistic: View Download Statistics

Actions (login required)

View Item View Item