Extensions of Modern Control and Estimation Methods to a Military Drive System

This paper investigates the application of Linear Quadratic Gaussian control to a nonlinear, military drive system. The major objective of this design is to eliminate the need of an expensive sensor used for disturbance measurement and, if possible, improve the system performance. Special attention is paid towards selecting efficient, accurate, control and estimator plant models. The resulting controller consists of a four state Linear Quadradic Regulator coupled with a eight state extended Kalman Filter, in addition to a nonlinear compensator to reduce multiplicative perturbations. The combined system was designed and evaluated via a detailed simulation. Baseline and sensitivity results indicate that the "modern" controller performs as well as the classically designed system, but without the need for an expensive sensor used for disturbance measurement.