未建模系统基于观测值的实时分块Kalman滤波估计方法研究

 本文以一类具有周期随机变化特点的随机过程为对象,在仅有测量模型的情况下研究估计方法的设计问题.首先,通过离散化方法建立点采样的离散输出方程、分块形式的输出方程以及描述点采样与被估状态块向量之间关系的输出方程;其次,利用待估变量具有的周期性随机游走特性,建立对应的状态模型;再者,利用扩展强跟踪滤波算法,分别得到了实时点估计滤波器、半实时块估计滤波器和实时块估计滤波器等三种未建模系统随机变量基于输出测量值的估计方法;最后,利用计算机仿真对三种滤波器的性能进行了比较分析.

Research on Real-Time Block Kalman Filtering Estimation Methods for the Un-modeled System Based on Output Measurements

Aiming at the random process with periodic changing characteristic,A new estimation method under the case having only measurement model is proposed in this paper.Firstly,the discrete output equation,output equation with blocking form,and output equation between point sample and the block vector of estimated state are taken.Secondly,by using the periodic random walk characteristic of estimated variable,the state models are established.Thirdly,based on strong tracking filter algorithm,three estimation methods such as real time point filter,semi-real time block filter,and real time block filter are proposed for the un-modeled random variable system with only output measurement.Finally,computer simulation is demonstrated to compare the performances of three proposed filters.