动态系统预报的一种新方法

本文给出了关于动态系统预报的一种新方法.这种方法把状态预报分成两个部分,即动 态系统参数的预报和以参数的预报值为基础的状态预报.误差分析与数值实例说明该方法可 以大大提高预报精度.

A New Method of Dynamic System Prediction

In this paper, we obtain a new method of dynamic system prediction. The predictor is divided into two parts, the prediction of time-varying parameters and the prediction of state or output of system. For the estimation of time-varying parameter θ(k), we applied "extension recursive gradient method". This method can be applied to both linear and non-linear cases. According to the estimation value sequence {θ(k)}, we obtain prediction estimation value {θ*(k + h)} ofθ(k). From this we be get h-step ahead prediction of output y(k). Computation results and error analysis are given to the effectiveness of this method.