卡尔曼滤波中动态模型的分级解耦

本文对卡尔曼滤波中动态模型的解耦提一种新的方法，这种方法其实质是把对整个动态模型的并联滤波估计变成对整个动态模型的分级串联滤波估计。这种分解耦的方法即可用于线性动态模型又可用于非线性动态模型。对于阶数很高的非线性动态模型，这种方法对于提高估计精度，减少量和计算时间，更有其优越性。本文利用一个九阶非线性动态模型进行仿真计算。计算结果说明了这种方法的有效性。

THE SERIAL DECOUPLING OF DYNAMIC MODEL IN KALMAN FILTER

This paper presents a new method of the dynamic model decoupling in kalman filter. This method change the parallel filter of whole model into serial filter of sub-model. It can be used not only in linear dynamic model, but also in nonlinear dynamic model. The method have more advantage in reducing the calculation time and improving the filter accuracy under the circum stance of high order dynamic model. The digital simulation of nine order nonlinear dynamic model demonstrates the effect of the method.