An implicitly adaptive model reference control design for linear multi-input multi-output systems

Abstract

The performance of the closed-loop systems is evaluated both analytically and experimentally using a computer simulation of the lateral dynamics of the F-8 aircraft and the control algorithm. Methods are developed for adjusting the closed loop performance of both the error response and the gain adaptation response. Using these methods, the performance of several exampies of the F-8 lateral dynamicsis adjusted to reduce any transient error between the plant and the mode

Since two of the sufficient conditions involve matrix equalities and inequalities, they are difficult to satisfy over a wide range of plant parameters. Thus, methods are proposed for reducing these sufficient conditions to a simpler set of algebraic inequalities using the positive real lemmanand a modified Riccati equation. These methods are demonstrated on a well formed model reference control problem involving the F-8 aircraft's lateral dynamics.

An adaptive, multi-input, multi-output, linear model reference control algorithm which does not require any parameter identification has been developed. This algorithm guarantees a bounded error between the plant and model if the parameters of the controller and the plant satisfy certain sufficient conditions. Unlike previous algorithms for multi-input, multi-output linear plants, these sufficient conditions do not include a modelmatching constraint. However,if the modelmatching constraint is satisfied, the error asymptotically approaches zero as with previousalgorithms. Thus, this algorithm is applicable to a wider spectrum of multi-input, ·multi-output, linear plants.