An order insensitive sequential fast covariance intersection fusion algorithm

Distributed fusion cubature Kalman filters for nonlinear systems

“…Therefore, the cross‐covariance matrices are indispensable when the distributed fusion estimator weighted by matrices is used. However, the calculation of cross‐covariance matrices is quite complex or even impossible for many practical systems, particularly for nonlinear systems. Next, we will give a general method to solve the cross‐covariance matrices between any two local estimators for multisensor nonlinear systems.…”

“…Therefore, the cross-covariance matrices P (ij) k|k are indispensable when the distributed fusion estimator weighted by matrices is used. However, the calculation of cross-covariance matrices is quite complex or even impossible for many practical systems, 33,34 particularly for nonlinear systems.…”

On equivalence of major relaxation methods for minimum ellipsoid covering intersection of ellipsoids

“…In [13], the authors also provide a suboptimal solution to the problem of finding the ellipsoid with minimum volume containing the intersection of the two ellipsoids defining the convex combination. In distributed fusion setting [14,15,16,17], when the cross-correlation of local sensor estimation errors is unknown or impractical, the covariance intersection (CI) algorithm are derived to deal with this problem in [18,19,20]. It provides not only a fused estimate point but also estimate covariance, the results are parameterized as convex combination of the local estimates.…”

Distributed two‐stage state estimation with event‐triggered strategy for multirate sensor networks

“…Some fusion algorithms with unknown cross-covariances were then developed such as the batch covariance intersection fuser, 5,6 the fast sequential covariance intersection fuser, 7 and the modified sequential fast covariance intersection fuser. In the work of Sun and Deng, 4 a multisensor optimal information fusion distributed Kalman filter weighted by matrices with a two-layer fusion structure was proposed, in which the variance and cross-covariance matrices are necessary.…”

“…However, in practice, the cross-covariances are usually unknown, or their computations are very complex especially for large-scale WSNs. Some fusion algorithms with unknown cross-covariances were then developed such as the batch covariance intersection fuser, 5,6 the fast sequential covariance intersection fuser, 7 and the modified sequential fast covariance intersection fuser. 8 These algorithms are based on the strongly connected WSNs.…”