应用现代时间序列分析方法 ,基于 ARMA新息模型和白噪声估值器 ,由非递推状态估值器的递推变形 ,提出了广义系统的 ARMA稳态最优递推状态估值器 .它们具有Wiener滤波器形式 ,可处理带奇异状态转移阵和 /或带相关噪声的广义系统 ,可统一处理滤波、平滑和预报问题 ,且可统一处理广义和非广义系统状态估计问题 .仿真例子说明了其有效性 . Based on ARMA innovation model and white noise estimator, the ARMA steady-state optimal recursion estimators of generalized systems are proposed based on recursive deformation of non-recursive state estimators using modern time series analysis methods. The Wiener filter, which deals with singular state transitions and / or associated noise, can deal with filtering, smoothing and forecasting problems uniformly and can handle the state estimation of generalized and non-singular systems in a unified way. Its effectiveness.