Closed loop filtered mrac for fast adaptation and low frequency learning with enhanced time delay margin

Major Problems of adaptive controllers for most of the safety critical systems such as aircraft, is the lack of robustness to time varying matched uncertainty, time delay and slow adaptation. In this paper a modified closed loop filtered MRAC architecture is presented that provides guaranteed transient, steady state performance, with fast adaptation and improved robustness to time varying matched uncertainty and time delay. In the proposed controller architecture tracking error is fed back to the reference model and projection based adaptive law is used to ensure uniform boundedness of the adaptive weight estimates. A low pass filter is used in the feedback loop to ensure that the control signal is in low frequency range. Some design guidelines are provided for designing of low pass filter based on small gain theorem. Illustrative numerical simulations verifies the effectiveness of the proposed work.