Genetic Design of Multivariable Flight-Control Systems Using Eigenstructure Assignment

Abstract

In the design of multivariable flight-control systems, performance specifications can frequently be met by assigning appropriate closed-loop eigenstructure by means of linear feedback control. However, although various powerful algorithms for the solution of this problem of eigenstructure assignment have been proposed, no universally applicable and computationally simple design methodology has yet emerged. The use of genetic algorithms to solve this fundamental problem of eigenstructure assignment is accordingly described, and illustrated by application to the design of multivariable fright-control systems for the AFTI/F-16 aircraft.