where ~xðtÞ is the relative estimation error and ^xðtÞ is a
vector of the estimated process states. The weighting
matrices Qobs 0 and Robs40 determine the convergence
properties of the estimate.
It is a well-established fact that the estimator feedback
matrix K that yields the optimal solution for
Equation (12) can be obtained by exploiting the duality
properties of the system (Bar-Shalom et al., 2001). In
essence, the observability of the pair ðA,CÞ is dual to
the controllability of the pair AT,CT
. Hence, by solving
the ARE for Equation (12) with AT,CT
, the optimal
observer gain is obtained as
where ~xðtÞ is the relative estimation error and ^xðtÞ is avector of the estimated process states. The weightingmatrices Qobs 0 and Robs40 determine the convergenceproperties of the estimate.It is a well-established fact that the estimator feedbackmatrix K that yields the optimal solution forEquation (12) can be obtained by exploiting the dualityproperties of the system (Bar-Shalom et al., 2001). Inessence, the observability of the pair ðA,CÞ is dual tothe controllability of the pair AT,CT . Hence, by solvingthe ARE for Equation (12) with AT,CT , the optimalobserver gain is obtained as
การแปล กรุณารอสักครู่..
