Adaptive control of arbitrary nonlinear inputs with feedforward compensation

This paper deals with Model Reference Adaptive Control (MRAC) with arbitrary bounded nonlinear inputs. A new analysis and design method is proposed. The main technique we used is feedforward compensation, with it the input nonlinearities may be partly cancelled by the feedforward controller and the closed-loop can be locally stabilized. Only standard assumptions of robust adaptive control is needed. We show that the modified scheme has the same stability and robustness properties as the standard one, but the performance are improved.