A new generation of adaptive controllers for linear systems

An approach to adaptive control of linear systems is presented. It foregoes the traditional certainty-equivalence concept, and treats the control of linear plants with unknown parameters as a nonlinear problem. The states of the resulting adaptive system converge to a manifold whose dimensions are smaller than with any previous scheme. The dynamic order of the resulting controllers is not higher (and in most cases is lower) than that of the existing adaptive schemes. A simulation comparison with a standard indirect linear scheme shows that the new nonlinear scheme achieves far superior transient performance without an increase in control effort.<<ETX>>