Optimal combination of plant and observation control using interpolation of some plant coordinates

The paper is concerned with optimal combination of plant motion and coordinate observation control with incomplete data. The plant coordinate and measurement evolution is described by linear vector-valued stochastic difference equations and the plant motion is repetitive. Conditional Gaussian filtering and interpolation reduce the problem to that of control with incomplete data which is decomposed into a problem of optimal plant control and a problem of observation optimization. An explicit form of optimal plant control is found. A deterministic problem is formulated for finding an optimal observation control. An example is given.