Modified strong tracking unscented Kalman filter for nonlinear state estimation with process model uncertainty

Abstract

This paper presents a modified strong tracking unscented Kalman filter (MSTUKF) to address the performance degradation and divergence of the unscented Kalman filter because of process model uncertainty. The MSTUKF adopts the hypothesis testing method to identify process model uncertainty and further introduces a defined suboptimal fading factor into the prediction covariance to decrease the weight of the prior knowledge on filtering solution. The MSTUKF not only corrects the state estimation in the occurrence of process model uncertainty but also avoids the loss of precision for the state estimation in the absence of process model uncertainty. Further, it does not require the cumbersome evaluation of Jacobian matrix involved in the calculation of the suboptimal fading factor. Experimental results and comparison analysis demonstrate the effectiveness of the proposed MSTUKF.

State Estimation for Sampled‐Data Descriptor Nonlinear System: A Strong Tracking Unscented Kalman Filter Approach

“…Based on the results in [18], a new approach of state estimation for sampled-data descriptor systems is proposed in this paper. The proposed sampled-data descriptor system is firstly discretized to obtain a discrete-time nonsingular model.…”

“…Recently, a strong tracking unscented Kalman filter (STUKF) method was proposed in [18] to investigate state estimation for discrete-time systems. In the presence of process model uncertainty, a defined suboptimal fading factor is introduced into the prediction covariance to adjust the Kalman gain matrix online.…”

Tracking a Decentralized Linear Trajectory in an Intermittent Observation Environment

“…Generally, a recursive Bayesian estimation algorithm sequentially updates current state estimates based on its previous estimate and the incoming data from the sensors, it is also called a Minimum Mean-Squared Error (MMSE) estimator. The Kalman filter is also considered as one of the most popular recursive Bayesian estimation algorithms and has been successfully used in various applications, including all forms of nuclear power plant instrumentation, navigation (aerospace, land, and marine), manufacturing, demographic modeling, the detection of underground radioactivity, and neural network training [5,10].…”

Estimation of Rotor Position in Brushless Direct Current Motor by Memory Attenuated Extended Kalman Filter

“…There are gaps of the brushless DC motor, for example, it is not easy to measure relevant parameters, the motor runs unstably, and there are errors in the motor model, when using a Kalman filter, it is prone to filter divergence, poor filter precision, and slow filtering. The MAEKF, if available, can ensure the filter convergence and stability by slipping old measured values from its memory but using new values [12][13][14].…”