Overcoming the Detectability Obstacle in Adaptive Output Feedback Control of DC-DC Boost Converter With Unknown Load

This brief proposes a solution to the long-standing problem of designing an adaptive control for the dc-dc boost converter with unknown load and unavailable input current measurement. The difficulty lies in the parametric uncertainty of the output dynamics, which poses a manifold of equilibria in the classical adaptive observer design. This is known as the detectability obstacle that imposes a restrictive assumption on the system behavior to ensure convergence. To overcome this problem, a class of saturated dynamic controllers is designed to guarantee the asymptotic regulation of the output voltage. An immersion- and invariance-based adaptive observer is proposed, which preserves the convergence property in conjunction with the controller with no need for an extra persistent excitation condition. To evaluate the transient behavior of the proposed controller, realistic simulations are provided, and then the performance comparison with two other well-known output feedback controllers is presented. Moreover, experimental results are concluded on a prototype dc-dc boost converter to support and verify the results of theory and simulations.