Abstract
An alternate approach to the design of gain-scheduled controllers is given. Standard gain-scheduling typically relies on the instantaneous value of the scheduling variable to update the control gains. This approach has the drawbacks of being limited to slow transitions between operating points and requiring a number of possibly tedious point-by-point designs. The present approach is to have the controller gains dynamically evolve according to the scheduling variable's history, rather than its instantaneous value. At the cost of increased order of the controller, this approach allows rapid transitions between operating conditions and alleviates the need for several point-by-point designs. The approach is demonstrated via a stabilizing control design for a longitudinal missile model.< >