Development of an algorithm for tuning a suboptimal scaling factor in the SINS filtering problem

2020 ◽  
Keyword(s):  
Process Model ◽  
Unscented Kalman Filter ◽  
Nonlinear Dynamical Systems ◽  
Scaling Factor ◽  
Fuzzy Filter ◽  
Nonlinear Dynamical ◽  
Sugeno Model ◽  
Takagi Sugeno ◽  
Softening Coefficient ◽  
Filtering Problem

A modified strong tracking unscented Kalman filter for nonlinear dynamical systems is proposed. A matrix of the suboptimal scaling factor is introduced into the prediction covariance to ensure evaluation stability and smoothness at appearance of the process model uncertainty. It is shown that the use of a fuzzy algorithm to adjust the softening coefficient in real time allows to avoid the loss of accuracy in the segments in which the process model is defined. As a result of modeling the SINS correction task, it was found that the proposed fuzzy filter has good evaluation smoothness and high accuracy. Keywords SINS; suboptimal scaling factor; softening coefficient; fuzzy Takagi — Sugeno model

Application of a Fuzzy Model to the Task of Filtering in Nonlinear Dynamic Systems

2020 ◽  
pp. 85-100
Author(s):  
N.P. Demenkov ◽  
D.M. Tran
Keyword(s):  
Kalman Filter ◽  
Dynamic Systems ◽  
Nonlinear Dynamic ◽  
Process Model ◽  
Unscented Kalman Filter ◽  
Fuzzy Model ◽  
Nonlinear Dynamic Systems ◽  
S Model ◽  
Takagi Sugeno ◽  
Softening Coefficient

In this paper, we consider various approaches to the problem of filtering in nonlinear dynamic systems and their algorithms. The Strong Tracking Unscented Kalman Filter, based on the combination of Unscented Kalman Filter and Strong Tracking Kalman Filter, provides stability to the uncertainty of the process model directly using a suboptimal scaling factor (SSF). The softening coefficient is part of the SSF and it improves the smoothness of the system state assessment. The coefficient is determined empirically and is included in the entire filtering process, which leads to a loss of accuracy in the time segments in which the process model is defined. The paper explores the use of Takagi --- Sugeno fuzzy model (T-S model) to adjust in real time the softening coefficient when the object's dynamics changes. As a result of a comparative analysis of the accuracy of the studied filters for the nonlinear model, it was found that the new filter using a fuzzy logical adaptive system possesses good smoothness of assessment and the greatest accuracy

THE UNSCENTED KALMAN FILTER, A POWERFUL TOOL FOR DATA ANALYSIS

2004 ◽  
Vol 14 (06) ◽  
pp. 2093-2105 ◽  
Author(s):  
A. SITZ ◽  
U. SCHWARZ ◽  
J. KURTHS
Keyword(s):  
Kalman Filter ◽  
Data Analysis ◽  
Dynamical Systems ◽  
Unscented Kalman Filter ◽  
Nonlinear Dynamical Systems ◽  
Simultaneous Estimation ◽  
Bayesian Filter ◽  
Nonlinear Dynamical

We present a derivation of the unscented Kalman filter (UKF) as an approximation to the optimal Bayesian filter equations. The potentials of the UKF are then demonstrated for the problem of simultaneous estimation of states and parameters from noise corrupted data of nonlinear dynamical systems. The UKF even works for the chaotic Chua system which includes nondifferentiable terms.

Mixed Unscented Kalman Filter and Differential Evolution for Parameter Identification

2012 ◽  
Vol 256-259 ◽  
pp. 2347-2353
Author(s):  
Ari Legowo ◽  
Zahratu H. Mohamad ◽  
Hoon Cheol Park
Keyword(s):  
Kalman Filter ◽  
Differential Evolution ◽  
Unscented Kalman Filter ◽  
Nonlinear Dynamical Systems ◽  
Spray Coating ◽  
Parameters Estimation ◽  
Effluent Treatment ◽  
Nonlinear Dynamical ◽  
De Algorithm ◽  
Tank System

This paper presents parameters estimation techniques for coupled industrial tanks using the mixed Unscented Kalman Filter (UKF) and Differential Evolution (DE) method. UKF have known to be a typical estimation technique used to estimate the state vectors and parameters of nonlinear dynamical systems and DE is one of the most powerful stochastic real-parameter optimization algorithms. Meanwhile, liquid tank systems play important role in industrial application such as in food processing, beverage, dairy, filtration, effluent treatment, pharmaceutical industry, water purification system, industrial chemical processing and spray coating. The aim is to model the coupled tank system using mixed UKF and DE method to estimate the parameters of the tank. First, a non-linear mathematical model is developed. Next, its parameters are identified using mixed Unscented Kalman Filter (UKF) and Differential Evolution (DE) based on the experimental data. DE algorithm is integrated into the UKF algorithm to optimize the Kalman gain obtained. The obtained results demonstrate good performances.

Sign in / Sign up

Export Citation Format

Share Document