Comparison of centralised scaled unscented Kalman filter and extended Kalman filter for multisensor data fusion architectures

Abstract

This study presents three non-linear centralised scaled unscented Kalman filter (SUKF) for multisensor data fusion algorithms, which are augmented measurements, measurements weighted and sequential filtering fusion. First, the accuracy analysis of extended Kalman filter (EKF) and SUKF is investigated in detail. Second, through comparing the error covariance traces and the absolute mean estimation errors of X and Y directions of centralised SUKF for multisensor data fusion algorithms with that of centralised EKF for multisensor data fusion algorithms, it can be remarked that the performance of centralised augmented measurements SUKF for multisensor data fusion algorithm is the best one among the six algorithms, which is to say that Algorithm (Iu) shows the best performance in accuracy. Finally, combining and synthetically analysing the running time of six algorithms, it illustrates that Algorithm (Iu) is optimal in comprehensive aspects among six algorithms.

Centralized Fusion Based on Interacting Multiple Model and Adaptive Kalman Filter for Target Tracking in Underwater Acoustic Sensor Networks

“…The Process Derivation of optimal CFKF Algorithm: If all the measurements are reliable, by using the projection theory [56], the innovation sequencez(k|k − 1) is defined as followsz (k|k − 1) = z(k) −ẑ(k|k − 1) (44) whereẑ (k|k − 1) = C(k)X c (k|k − 1)…”

“…The combined tracks at local platforms were transmitted to the fusion center and further fused there with a constructed global model [43]. There are basically two fusion architectures: centralized [44], [45] and distributed [46]- [48]. For the distributed fusion, which is also called as the state-vector or track fusion, a group of local Kalman filters are used in parallel to obtain individual sensor-based estimates and the distributed fusion formulae are then applied to yield an improved joint estimate.…”

Improved square root adaptive cubature Kalman filter

“…Owing to the practicability and flexibility, non-linear filters have been widely studied in recent decades [1][2][3][4][5][6][7][8][9][10][11][12]. Motivated by the solution of intractable numerical integrals within the filter framework, various non-linear filters are proposed.…”

“…Motivated by the solution of intractable numerical integrals within the filter framework, various non-linear filters are proposed. The typical filters are the extended Kalman filter (EKF) [1,2], unscented Kalman filter (UKF) [3][4][5], Gauss-Hermite quadrature filter (GHQF) [6,7], sparse grid quadrature filter (SGQF) [8][9][10] and cubature Kalman filter (CKF) [11,12].…”

Research into a Multi-Variate Surveillance Data Fusion Processing Algorithm

“…In recent years, the KF (Kalman Filter) [2], MKF (Mixture Kalman Filter) [3], EKF (Extended Kalman Filter) [4,5] have been used for multi data fusion. Ma and Wang proposed an RFID (Radio Frequency Identification) tracking method which combines the received signal strength with phase shift to predict the instantaneous position of a moving target.…”