Backstepping with virtual filtered command: Application to a 2D autonomous Vehicle

Abstract

Through this work a deep understanding of the backstepping control technique is sought when applied over non-affine systems. It is shown that in this case appears the necessity to bound the value of internal states and that a modification over standard backstepping is mandatory. The principal goal of this study is to evaluate the effects of finite frequency filters, and the effects of saturation affecting intermediate states and control actions, in the tracking performance when using the command filtered backstepping. Some relations that bind the controller gains to maintain performance appear naturally. Finally simulations over a 2D steering robot model are given to illustrate the found results.