Minimum Expected Cost Control of a Remotely Piloted Vehicle

Abstract

A technique for designing constant gain feedback controllers for linear systems having uncertain or variable parameters is presented and demonstrated for a realistic design of a remotely piloted vehicle. This controller design technique—named Minimum Expected Cost Control—produces closed-loop system behavior which is acceptable for all values of the parameters within specified limits and is optimum in some overall sense. Only a small number of points in parameter space are explicitly considered in the design algorithm, making it computationally fast and inexpensive. The technique is used to design a constant-gain lateral autopilot for a remotely piloted vehicle that will fly at a wide range of altitudes and airspeeds. Both full and partial state feedback situations are considered.