New method for optimal control and filtering of weakly coupled linear discrete stochastic systems

In this paper the algebraic regulator and filter Riccati equations of weakly coupled discrete-time stochastic linear control systems are completely and exactly decomposed into reduced-order continuous-time algebraic Riccati equations corresponding to the subsystems. That is the exact solution of the global discrete algebraic Riccati equation is found in terms of the reduced-order subsystem nonsymmetric continuous-time algebraic Riccati equations. In addition, the optimal global Kalman filter is decomposed into local optimal filters both driven by the system measurements and the system optimal control inputs. As a result, the optimal linear-quadratic Gaussian control problem for weakly coupled linear discrete systems takes the complete decomposition and parallelism between subsystem filters and controllers.<
>