Combined UKF/KF for Fast In-Motion Attitude Determination of SINS

Trajectory tracking of high speed maneuvering target in air by composite particle filter

“…Filtering is based on the target observation information, combined with the model information, to obtain the optimal estimate of the system state.Filtering algorithms represented by Kalman Filter (KF) are often used to solve the problem of target trajectory tracking. KF algorithm is only suitable for linear Gaussian system, while EKF and Cubature Kalman Filter are often used for complex nonlinear systems.CKF, Unscented Kalman Filter (UKF) and Particle Filter (PF).The so-called nonlinear filtering algorithm, most of the nonlinear system through a certain way to "linear" fitting, the process of linearization will lead to a large deviation of the system state estimator, linear/nonlinear fusion filtering method provides a solution.In reference [16] , a combined (UKF/KF) structured attitude determination method is proposed. Firstly, UKF is used to reduce the initial attitude error to a small range during rough alignment, and then KF filtering is used to further reduce the fine alignment of attitude error.In reference [17] , KF-UKF signal processing algorithm was proposed to estimate the position and speed of maglev train.Zhao Yanming et al [18] proposed the KF/EKF2 hybrid filter to solve the problem that the state dimension of nonlinear filter is too large under the large error Angle of SINS.For the tracking problem of high speed maneuvering target, most of the existing studies use single nonlinear filtering, and the tracking accuracy of target motion state is insufficient.…”