Neuro-Adaptive Output Feedback Control for a Class of Nonlinear Non-Minimum Phase Systems

摘要

This paper presents an adaptive output-feedback control method for non-affine nonlinear non-minimum phase systems that have partially known Lipschitz continuous functions in their arguments. The proposed controller is comprised of a linear, a neuro-adaptive and an adaptive robustifying control term. The adaptation law for the neural network weights is obtained using the Lyapunov's direct method. One of the main advantageous of the proposed method is that the control law does not depend on the state estimation. This task is accomplished by introducing a strictly positive-real augmented error dynamic and using the Leftshetz-Kalman-Yakobuvich lemma. The ultimate boundedness of the error signals will be shown analytically using the extension of Lyapunov theory. The effectiveness of the proposed scheme will be shown in simulations for the benchmark problem Translational Oscillator/Rotational Actuator (TORA) system.