Consensus Based Strong Tracking Adaptive Cubature Kalman Filtering for Nonlinear System Distributed Estimation
Zheng Zhang,Qingdong Li,Liang Han,Xiwang Dong,Yuan Liang,Zhang Ren
DOI: https://doi.org/10.1109/access.2019.2929302
IF: 3.9
2019-01-01
IEEE Access
Abstract:The distributed cubature Kalman filter is widely used in the field of target tracking, however, the presence of model uncertainties will undermine its tracking stability and effectiveness for tracking maneuvering target. In order to eliminate this effect on maneuvering target tracking, this paper develops a distributed consensus based strong tracking cubature Kalman filter scheme. First, each node obtains its local estimation with the usage of local observations via strong tracking cubature Kalman filter, where the suboptimal fading factor and adaptive factor are introduced for adaptively modifying the filter gain. Then, the designed filter gain is used for updating the local state estimation. Second, after all, nodes have achieved its local estimation, each node exchanges its local estimation to its neighbors and updates its local estimation according to the consensus communication protocol. It can be further proved that the distributed interaction between neighbors will contribute to enhancing the tracking stability. The detailed proof for stochastic boundedness of the estimation error is analyzed by introducing a stochastic process. Simulation results demonstrate that the proposed algorithm can achieve higher tracking accuracy than the existing methods for tracking a maneuvering target.