Design of signal-subspace cost functionals for parameter estimation
Wenyuan Xu, M. Kaveh · 1992
A probabilistic approach to the quantification of the resolving ability of a general class of MUSIC type estimators (m-estimators) is presented. Based on a resolution-maximizing criterion of optimality, a cost functional is designed for a specific parametric subclass of m-estimators. An effective data-adaptive value for the parametric class is also presented and the results are generalized to a broader nonparametric subclass.>