状态估计在最小方差控制中的应用

The Application of State Estimation in Minimum Variance Control

通过引入状态变量,将普通的控制系统改写为状态空间模型,应用时域上的现代时间序列分析方法和射影理论,基于ARMA新息模型和白噪声估值器,给出状态变量的滤波估值,并用状态变量的滤波代替控制系统中的观测,从而得到含有新息补偿的最小方差控制律,它比直接由控制系统得到的最小方差控制律更加准确,也更具应用价值。

A normal control system can be overwritten as a state-space model by the introduction of state variables. By the modern time series analysis method and projective theory, based on the autoregressive moving average （ARMA） innovation model and white noise estimator, the state filter can be obtained. A minimal variance control law with innovation compensation was given by the state filter instead of control system＇s observer. It is more accurate than the minimum variance control law of direct control system, and has more application value too.