Feedback and Feedforward LQG Controller Design

A solution for the LQG stochastic optimal control problem is obtained in the z-domain for systems where both feedforward and feedback control may be used. A two degrees of freedom feedback structure is employed and the system includes both coloured measurement noise and both measurable and unmeasurable disturbances The solution is obtained in a simple form which is particularly appropriate in self-tuning systems. This involves the solution of diophantine equations, associated with each controller input (from the feedback, reference and feedforward signals) and the calculation of spectral factors.