Optimal control of linear stochastic system using smoothed estimate of phase coordinates

Consideration was given to the problem of optimizing the control of a linear stochastic plant in the linearly quadratic performance index. The optimality criterion is represented by the minimal root-mean-square form of deviation of the plant state vector from some desired vector time function under a limited root-mean-square form of the control vector. The required time function represents the optimal linear estimate of smoothing with constant delay of the state vector of another stochastic plant which is based on imprecise observations. The differential equations of the optimal information and control algorithm were established. A particular case of the general problem formulation, the control of a pilotless flight vehicle in the mode of rounding relief was considered by way of example.

Presented by the member of Editorial Board: A. I. Kibzun