Optimal output feedback using dyadic decomposition

The literature is replete with methods to numerically compute the gain matrix for optimal output feedback controllers. Inherent to the vast majority of these methods is the solution of necessary conditions for optimality. This work presents a fundamentally different numerical scheme for computing H/sub 2/ optimal output feedback controllers. The method is based on a direct and analytic solution to the cost function using dyadic decomposition. Analytic solutions for the first and second derivatives of the cost function with respect to gain matrix elements are written in compact form allowing efficient solution using robust second order Newton-Raphson optimization. Using a flight control example, the method is exercised and compared to other optimal output feedback methods.