Robust Adaptive Control with a Modified Controller for Transient Response Improvement

Although rigorous mathematical results have been obtained to guarantee the steady-state convergence properties of adaptive control systems, the potential poor transient response imposes limitations on the practical application of adaptive control schemes. To achieve fast adaptation, a high-gain learning rate is usually used to suppress uncertainties rapidly, while a high-gain learning rate can create high-frequency oscillations in the control signals, and then trigger system instability. In this paper, we construct a new model reference adaptive controller (MRAC) for a class of nonlinear uncertain dynamic systems to improve steady-state and transient performance simultaneously. The proposed controller involves an extra compensation term which can capture the unknown system dynamics effectively. By using this modified controller to reshape the controlled system, it is shown in simulations that high-frequency components in the tracking error can be diminished. In this case, we can use a high-gain learning rate to achieve fast adaptation without exciting high-frequency unmodeled dynamics. Extensive simulation results show improved control response of the proposed control schemes in comparison to the standard MRAC.