Resolution of regularized output least squares estimation procedures

摘要

The regularized output least squares estimation technique is considered as an example of an estimation procedure in order to evaluate sensitivity of mappings that arise in the context of identical twin test problems. For such test problems, one specifies a reasonable test parameter to generate data from which the test parameter is to be recovered by means of an estimation procedure. The properties of the mapping taking the test coefficient to the optimal estimator are studied. Conditions are determined under which the mapping is well-defined as a function of the set of admissible parameters into itself. Resolution properties are related to the properties of the derivative of this mapping. Examples are considered using this information to formulate what is meant to be an optimal observer and to determine the optimal location of observers.