非线性离散随机时变系统比例微分滤波

Proportion-differentional Filtering of Nonlinear Discrete Time-varing Stochastic Systems

扩展卡尔曼滤波 (EKF)是从极小化状态估计误差的方差得到的 ,没有考虑状态误差的变化率 ,因而对非线性时变系统EKF估计方法惯性作用较大 ,从而产生估计滞后。提出了非线性离散随机系统比例微分滤波 (PDF) ,PDF联合考虑极小化状态估计误差方差和状态误差变化率的方差 ,克服了EKF对非线性时变系统估计滞后的缺点 ,估计具有适时性 ,提高了估计的精度。仿真例子证明了所提出的估计方法的有效性。。

Extended Kalman Filtering (EKF) was derived from minimizing variance of state estimation error. Not considering its rate of change, EKF method has large inertia for nonlinear time-varing systems and resulted in estimation lag. Proportion-Differential Filtering(PDF) of nonlinear discrete time-varing stochastic systems is presented. PDF is derived not only from taking into account minimizing variance of state estimation error, but regarding its rate of change, so the estimation method can update information in time , avoid estimation lag of EKF, and raise estimated accuration. A simulation example shows its effectiveness.