Recursive functional series modeling estimators for identification of time-varying plants-more bad news than good?

摘要: The properties of a class recursive estimators-the exponentially weighted functional series modeling estimators-are discussed. These estimators can be used, e.g. in adaptive prediction or control applications. It is argued that there exists relationship between the amount information about time-varying system parameters, which available priori, and robustness identification algorithm based on such prior knowledge. more specialized estimation is, less reliable it might under nonstandard conditions. This reason why simple algorithms as least squares have to recommended if no nonstationarity advance. >