Xiao Mengli, Zhang Yongbo, Wang Zhihua, Fu Huimin
Research Center of Small Sample Technology, Beihang University, Beijing 100191, China.
Research Center of Small Sample Technology, Beihang University, Beijing 100191, China.
ISA Trans. 2018 Apr;75:101-117. doi: 10.1016/j.isatra.2018.02.007. Epub 2018 Feb 20.
Considering the performances of conventional Kalman filter may seriously degrade when it suffers stochastic faults and unknown input, which is very common in engineering problems, a new type of adaptive three-stage extended Kalman filter (AThSEKF) is proposed to solve state and fault estimation in nonlinear discrete-time system under these conditions. The three-stage UV transformation and adaptive forgetting factor are introduced for derivation, and by comparing with the adaptive augmented state extended Kalman filter, it is proven to be uniformly asymptotically stable. Furthermore, the adaptive three-stage extended Kalman filter is applied to a two-dimensional radar tracking scenario to illustrate the effect, and the performance is compared with that of conventional three stage extended Kalman filter (ThSEKF) and the adaptive two-stage extended Kalman filter (ATEKF). The results show that the adaptive three-stage extended Kalman filter is more effective than these two filters when facing the nonlinear discrete-time systems with information of unknown inputs not perfectly known.
考虑到传统卡尔曼滤波器在遭受随机故障和未知输入时性能可能会严重下降，这在工程问题中非常常见，本文提出了一种新型自适应三阶段扩展卡尔曼滤波器（AThSEKF），以解决在这些条件下非线性离散时间系统中的状态和故障估计问题。推导过程引入了三阶段UV变换和自适应遗忘因子，并且通过与自适应增广状态扩展卡尔曼滤波器进行比较，证明其具有一致渐近稳定性。此外，将自适应三阶段扩展卡尔曼滤波器应用于二维雷达跟踪场景以说明效果，并将其性能与传统三阶段扩展卡尔曼滤波器（ThSEKF）和自适应两阶段扩展卡尔曼滤波器（ATEKF）的性能进行比较。结果表明，当面对未知输入信息不完全已知的非线性离散时间系统时，自适应三阶段扩展卡尔曼滤波器比这两种滤波器更有效。
