This paper describes an applications oriented approach to the generation of optimal output feedback gains for linear time-invariant systems which is dependent of the open loop stability.
The standard requirements for the provision of initial stabilizing output feedback gains for priming the computational process is circumvented. In lieu of initial stabilizing gains, the proposed algorithm employs the full state feedback solution which guarantees stability under the mild condition of stabilizability.
In this paper the generation of sub-optimal output feedback problem is cast in the setting of a constrained parameter optimization problem. The solution of this constrained optimization employs Hestenes' method of multipliers with some modifications. A primal-dual problem is considered where the primal minimization employs a Davidon-Fletcher-Powell method, and the dual maximization is accomplished via a quasi-Newton procedure.
This approach provides the designer with a means of suppressing to zero selected gains corresponding to accessible output, either for the purpose of simplifying the controller structure or because prior knowledge indicates that certain gains are "nonproductive". In addition, the designer can easily incorporate certain linear constraints, which the feedback gains will satisfy, into the proposed procedure.
Detailed algorithm description and computational results for a realistic flight control design problem are provided.