基于UKF的高斯和滤波算法

介绍了扩展卡尔曼滤波算法和无迹变换（unscented transformation，UT）算法，并对扩展卡尔曼滤波算法（EKF）和无迹卡尔曼滤波算法（UKF）进行比较，阐明了UKF优于EKF。在此基础上，提出了一种基于Unscented变换（UT）的高斯和滤波算法，该算法首先通过合并准则得到适当个数的混合高斯模型，逼近系统中非高斯噪声的概率密度；然后，再通过UT算法进行滤波。最后分别对基于EKF和UKF的滤波方法进行实验，并对实验结果进行比较与分析，验证了算法的有效性和优良性。

A Gaussian Sum Filter Based on UKF

This paper gives a brief introduction to EKF and unscented transformation, and makes a comparision between EKF and UKF, and shows the advantages of UKF over EKF. On this basis, it presents a gaussian sum filter based on unscented transformation. Firstly, the Gaussian Mixture Model with appropriate numbers is got by the combination norm, which is used for approximating to the density of non - gaussian noise, then a process of filtering based on EKF and UKF is implemented individually and an analysis of the experiment results is made. Finally, the validity and the advantages of this method are well verified hy the experiments.