ABSTRACT
This chapter considers nonlinear dynamic compensators design from yet another angle, as multiwindow composite controllers. Multiwindow control makes transitions between the elementary linear controllers on the basis of the amplitude of the error. The design problem, therefore, arises of integrating these locally optimal linear controllers into a composite nonlinear controller and providing smooth transitions between these regional linear laws. The multiwindow nonlinear controller can be implemented as follows: the error is partitioned into components falling into different windows, the components processed by the linear operators of the windows, and the results added up or combined by nonlinear functions. This architecture is referred to as multiwindow, and a great number of useful nonlinear control schemes can be cast in this form. A strong correlation exists in many systems between the error’s amplitude and the error frequency content, so that the errors fall into a set of diagonal windows.
