Minimum Sensitive Linear Feedback Compensators

In the design of optimal control systems, emphasis is placed on the accuracy of the system mathematical model. If certain modeling parameters deviate from their assumed nominal values, the optimal control may not produce the desired output. A complete theory is developed in the paper for the practical design of linear, nominally equivalent feedback compensators which minimize the output sensitivity to system parameter variations. An example is presented to compare these compensators with those which regulate and stabilize.