Abstract: This paper presents an in-depth survey on the Unscented Transformation（UT） and the Unscented Kalman Filtering（UKF） algorithm，including the origination of the nonlinear estimation，the framework and the basic algorithm of UT，various evolutions of UT，the design guideline of sigma-point set，the basic algorithm of UKF and its variants，the further insight into UT，the comparison between UT and several other nonlinear derivative-free filtering methods，the application areas pertinent to UT and UKF，and a simulation example based on several typical forms of UKF.In the last part of the paper，some still unresolved problems about UT and UKF are exposed and an outlook of its future developments is explicitly listed too.Above all，some newest achievements and opinions are included in this paper，which are based on the author’s own research findings.

Key words: nonlinear estimation, Unscented Transformation（UT）, Unscented Kalman Filtering（UKF）, extended Kalman filtering, particle filtering, statistical linear regression

摘要： 综述了非线性估计问题的由来、无味变换（UT，Unscented Transformation）的基本思路与基本算法、各种衍变形式、σ点集的设计原则、无味卡尔曼滤波（UKF，Unscented Kalman Filtering）的基本算法及其各种改进算法、UT的本质、UKF与几种免微分非线性滤波方法的比较、UT与UKF的相关应用、针对几种UKF算法的仿真实例，以及目前在UT与UKF的研究中尚存在的一些问题和对今后研究的展望等；提出了笔者的一些最新研究成果和见解。

关键词: 非线性估计, 无味变换, 无味卡尔曼滤波, 扩展卡尔曼滤波, 粒子滤波, 统计线性回归