一类结构适应式自校正预报器及其应用

A Structure-adaptive Self-tuning Predictor and Its Application

对ARMAX系统提出了一种线性结构适应式自校正预报器SASTP,它由最优(最小方差)预报器和变结构随机逼近算法组成.通过实例应用,对于一般的自校正预报器的70步跟踪预报,平均相对误差为2.42％.

A structure-adaptive self-tuning predictor is given to ARMAX system. It consists of optimal (minimum variance) predictors and stochastic approximation algorithms. Through living example, the average relative error is 2.42 percent to usual self-tuning predictor at 70 step following predut.