2006•Unpublished venueRequires access

An adaptive nonlinear filter of discrete-time system with uncertain covariance using unscented Kalman filter

Wanchun Li, Ping Wei, Xiao Xian-Ci

Open publisher page 7 citations

Abstract

A novel adaptive unscented Kalman nonlinear filter (AUKF) is presented in this paper. In many system, the noise covariance is unknown exact, but the approximate can been obtained by many methods. The approximate is used to initialize the unscented Kalman filter (UKF). Each step these noise covariance are adjusted based on the prior noise covariance and state information. To reduce obsolete measure value and covariance, a limited memory method is used. On the performance, UKF is better than EKF. The AUKF are better than these adaptive Kalman filters which based on extended Kalman filter (EKF). A target tracking is used to demonstrate this.

About this research paper

What this paper is about

A novel adaptive unscented Kalman nonlinear filter (AUKF) is presented in this paper. In many system, the noise covariance is unknown exact, but the approximate can been obtained by many methods. The approximate is used to initialize the unscented Kalman filter (UKF). Each step these noise covariance are adjusted based on the prior noise covariance and state information. To reduce obsolete measure value and covariance, a limited memory method is used. On the performance, UKF is better than EKF. The AUKF are better than these adaptive Kalman filters which based on extended Kalman filter (EKF). A target tracking is used to demonstrate this.

Why it matters

OpenAlex reports 7 citations for this work. Citation counts describe recorded attention and do not establish research quality.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

A novel adaptive unscented Kalman nonlinear filter (AUKF) is presented in this paper. In many system, the noise covariance is unknown exact, but the approximate can been obtained by many methods. The approximate is used to initialize the unscented Kalman filter (UKF). Each step these noise covariance are adjusted based on the prior noise covariance and state information. To reduce obsolete measure value and covariance, a limited memory method is used. On the performance, UKF is better than EKF. The AUKF are better than these adaptive Kalman filters which based on extended Kalman filter (EKF). A target tracking is used to demonstrate this.

Key concepts: Extended Kalman filter, Kalman filter, Control theory (sociology), Invariant extended Kalman filter, Unscented transform, Fast Kalman filter, Covariance intersection, Covariance

Related papers

Back to paper searchBrowse research topicsOriginal source
An adaptive nonlinear filter of discrete-time system with uncertain covariance using unscented Kalman filter — Research Paper | ScholarLens