Abstract
Filtering and estimation are two important tools of engineering. Whenever the state of the system needs to be estimated from the noisy sensor measurements, some kind of state estimator is used. If the dynamics of the system and observation model are linear under Gaussian conditions, the root mean squared error can be computed using the Kalman Filter. But practically, noise frequently enters the system as not strictly Gaussian. Therefore, the Kalman Filter does not necessarily provide the better estimate. Hence the estimation of the nonlinear system under non-Gaussian or quasi-Gaussian noise is of an acute interest. There are many versions of the Kalman filter such as the Extended Kalman filter, the Unscented Kalman filter, the Ensemble Kalman filter, the Particle filter, etc., each having their own disadvantages. In this thesis work I used a bridging strategy between the Ensemble Kalman filter and Particle filter called an Ensemble Kalman Particle filter. This filter works well in nonlinear system and non-Gaussian measurements as well. I analyzed this filter using MATLAB simulation and also applied Gaussian Noise, non-zero mean Gaussian Noise, quasi-Gaussian noise (with drift), random walk and Laplacian Noise. I applied these noises and compared the performances of the Particle filter and the Ensemble Kalman Particle filter in the presence of linear and nonlinear observations which leads to the conclusion that the Ensemble Kalman Particle filter yields the minimum error estimate. I also found the optimum value for the tuning parameter which is used to bridge the two filters using Monte Carlo Simulation.
Advisor
Vincent Winstead
Committee Member
Han-Way Huang
Committee Member
Namyong Lee
Date of Degree
2016
Language
english
Document Type
Thesis
Degree
Master of Science (MS)
Department
Electrical and Computer Engineering and Technology
College
Science, Engineering and Technology
Recommended Citation
Prabhakaran Jeyakumari, S. (2016). Use of Bridging Strategy between the Ensemble Kalman Filter and Particle Filter for the Measurements with Various Quasi-Gaussian Noise [Master’s thesis, Minnesota State University, Mankato]. Cornerstone: A Collection of Scholarly and Creative Works for Minnesota State University, Mankato. https://cornerstone.lib.mnsu.edu/etds/640/
Creative Commons License
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License
Included in
Computational Engineering Commons, Electrical and Computer Engineering Commons, Other Engineering Commons
Rights Statement
