Distributed Kalman filtering for uncertain dynamic systems with state constraints

Abstract

This article addresses the distributed state estimation problem for uncertain time-varying dynamic systems with state constraints over a sensor network. By using a null space method, the distributed state estimation problem for uncertain dynamic systems with state constraints can be cast into a new unconstrained distributed state estimation problem for reduced uncertain dynamic systems. A constrained distributed Kalman filter is proposed, and it is shown that the full state estimates can be recovered at any time and satisfy the constraints. An optimized upper bound of the estimation error covariance of each sensor is obtained, and the corresponding gains are designed. The application conditions of the proposed algorithm are mild, and they can be off-line checked. Furthermore, the computational requirements in this article are also reduced compared with the existing results. Finally, the performance of the proposed filter algorithm is demonstrated through numerical simulations.

Distributed maximum correntropy unscented Kalman filtering with state equality constraints

“…The predicted error covarianceΣ i k|k−1 may be a positive semidefinite matrix. In order to take the inverse operation, a very small positive constant is used in (19). IfΣ i k|k−1 is a positive definite matrix and…”

Distributed cubature Kalman filtering for nonlinear systems with stochastic communication protocol

“…In addition, for high nonlinear strength, the unignorable linearization errors would probably lead to poor filtering performance. Accordingly, some effective nonlinear filtering strategies have been discussed, such as in previous studies [5][6][7][8], among which the method of the cubature Kalman filtering (CKF) has been confirmed a powerful tool for the nonlinear systems to settle filtering problem.…”

Event‐triggered resilient distributed extended Kalman filter with consensus on estimation

“…As a popular method, consensus-based distributed filters can be mainly classified into the following three categories: (i) consensus on estimation (CE) filter has been considered by adding a consensus term with a predefined gain in the traditional Kalman filter. [2][3][4] The main drawback of the CE filter is that the covariance information is not employed efficiently in a distributed manner (ii) consensus on measurement (CM) filter in which the consensus is obtained on the local measurements and innovation covariances. However, it requires high communication costs and it does not assure convergence unless the number of consensus iterations are sufficiently large.…”