ROBUSTNESS AND PERFORMANCE OF A MODIFIED ADAPTIVE BACKSTEPPING CONTROLLER

Recently, a new class of adaptive control schemes based on non-linear design techniques, have been proposed for minimum phase linear time in variant plants, Under certain assumptions on the plant transfer functi on, these schemes guarantee uniform signal boundedness and good transi ent and steady-state response for the regulation or tracking error. In this paper we propose a modification that improves the robustness of these schemes with respect to a class of multiplicative uncertainties and input, output disturbances, without loss of performance. This lack of a trade-off between performance and robustness is a result of rest ricting the multiplicative unmodelled dynamics to be small in both the low- and high-frequency range, It is also shown that parameter conver gence and performance improvement is achieved with a dominantly rich r eference input. (C) 1998 John Wiley & Sons, Ltd.