Comments on two recent papers on reduced-order optimal state estimation for linear systems with partially noise corrupted measurement
J. O’Reilly · IEEE Transactions on Automatic Control · 1982
Two recent correspondence on the optimal singular state estimation problem are shown to be in error. Optimal (linear least-squares) estimators, based on an innovations approach [9], [10], are presented. An important outcome is the clarification and hopeful elimination of some current misconceptions in relation to stochastic observer theory.