NUMERICAL METHODS FOR OPTIMAL OBSERVATION OF NONLINEAR DYNAMICAL SYSTEMS

We consider a problem of state estimation for nonlinear continuous time system ssubject to set-membership uncertainty in initial states and measurement errors. The optimal observation problem is formulated and algorithms for its a posteriori and positional solution are presented. Local optimization involves solution of a linear approximation of the original problem and its correction by asymptotic technique. Global optimization uses the same idea but is based on piecewise linear approximation. The results are illustrated by examples.