Direct Uncertainty Minimization in Model Reference Adaptive Control: Experimental Results

摘要

This paper presents an application of a recently developed direct uncertainty minimization approach for model reference adaptive control on a dual-rotor helicopter testbed. Specifically, the direct minimization approach uses a gradient minimization procedure to construct modification terms that are included in the adaptive control law and the update laws to improve the closed-loop system performance. These modification terms are activated when the system error between the uncertain dynamical system and a given reference model capturing the desired closed-loop dynamical system behavior is nonzero and then vanish as the system reaches steady-state. The experimental results in this paper demonstrate the capability of the direct minimization approach to improve the closed-loop system performance and enforce performance bounds.