Robust Command Augmentation System Design Using Genetic Methods

摘要

This paper describes the use of a genetic search method in the design of a command augmentation system for a high-performance aircraft. A genetic algorithm is used in the design of H(infinity) controllers for the longitudinal and lateral-directional channels by selecting the weighting functions. The integral of absolute value of error between the actual response and that of an ideal model is used as the fitness criterion, along with additional terms to penalize for cross-coupling between Ps and ny; non-minimum phase behavior, and the closed-loop infinity-norm bound, gamma. Starting from an initial population of weighting functions, the algorithm generates new functions with the goal of improving the fitness. These controllers are then evaluated in a 6 degree-of- freedom nonlinear model of the aircraft.