基于非线性预测滤波和UKF的状态估计方法

非线性系统存在建模误差时,UKF的状态估计误差较大,为了提高UKF对非线性系统的状态估计能力,本文将非线性预测滤波(NPF)方法和UKF相结合,提出了一种改进的UKF。首先应用NPF求得模型误差值,得到非线性系统的修正模型,将模型离散化再应用UKF进行状态估计。在仿真实验中分别应用单纯的UKF和改进后的UKF对一个存在模型误差的非线性系统进行状态估计,对它们的估计结果进行了比较和分析,结果表明结合NPF的UKF能够提高非线性系统状态估计的精度。

State Estimation Based on Nonlinear Predictive and Unscented Kalman Filter

When modeling error exists, estimation error of UKF will be large. In order to improve the state estimation ability, an improved UKF is proposed in this paper, in which Nonlinear Predictive Filter （NPF） is integrated with Unscented Kalman Filter. At first, modeling error is calculated with NPF, so corrected model of nonlinear system is acquired who is diseretized then and UKF is applied to estimate its state. In simulation experiments standard UKF and improved UKF are used to estimate state of nonlinear system, estimation results indicate that improved UKF can enhance estimation precision of nonlinear system.