关键词: 非线性高斯滤波, 广义高阶容积准则, 自适应滤波, 目标跟踪

Abstract: In order to overcome the problem that Cubature Kalman Filter decreases in accuracy when system states suddenly change, combined with the generalized high-degree Cubature Kalman Filter and strong tracking filter algorithm, an Adaptive Generalized High-degree Cubature Kalman Filter（AGHCKF） is established. The generalized high-degree cubature rule and the diagonalization of matrix are used to improve the accuracy and the stability of filtering algorithm. Furthermore, the strong tracking filter algorithm is introduced to improve the capability of the filter to deal with uncertainty factors by modifying the predicted states’ error covariance with a fading factor and the residual sequence is forced to be orthogonal. A maneuvering target tracking problem with unknown sudden states changes in system states is used to test the performance of AGHCKF. The simulation results indicate that AGHCKF can achieve good filtering performance when states’ changes suddenly occur, with great robustness and better system adaptive capacity.

Key words: nolinear Gaussian filter, generalized high-degree cubature rule, adaptive filter, target tracking