一种基于状态方差阵对角相似分解的UKF 算法

在强非线性和复杂性的系统下，普通对称采样UKF 算法存在稳定性问题。为此，本文提出了基于状态方差阵对角相似分解的UKF 算法来保证算法的状态方差阵半正定。同普通对称采样UKF 算法比较，该算法降低了对状态方差阵的正定性要求。仿真试验验证了该方法的有效性。

The general symmetric sampling UKF (unscented Kalman filter) algorithm is unstable due to the strong nonlinearity and complexity of the system. Therefore, a novel UKF algorithm based on the diagonal similar decomposition of state covariance (DSDUKF) is proposed which can ensure the state covariance to be positive semidefinite. Comparing with the general UKF, the DSDUKF algorithm relaxes the condition of the positive semi-definiteness of state covariance. Simulation experiments demonstrate the effectiveness of the algorithm.

国家自然科学基金资助项目（60705028，60775056）